fix(server): make OpenCode health check resilient to terminal escapes and spaced model IDs

[Mostafa-Ben-Git]

Closes#8386

Problem

opencode <= 1.18 emits an OSC title sequence (ESC ]0;<cwd>: ready BEL) on stdout even when piped, and some custom providers use model IDs with spaces (e.g. Kimi K3). On Windows this caused:

• parseServerUrlFromOutput to miss the opencode server listening line,
• parseModelsCliOutput to drop the model with a spaced ID and its predecessor (regex ^(\S+\/\S+)\s*$ doesn't allow spaces after the slash),
• parseAgentListCliOutput / parseSkillsCliOutput to retain raw control bytes inside decoded strings,
• and loadInventoryFromCli to surface OpenCode models command exited with code 1. with no stderr/stdout, making the failure unactionable.

All three parser bugs have open PRs (#7834, #7988, #8302, #7755) but were still present on main.

Fix

Ported from those PRs and combined:

1. New helperpackages/shared/src/stripTerminalEscapes.ts — removes OSC (terminated by BEL or ST) and CSI sequences.
2. parseServerUrlFromOutput / parseModelsCliOutput / parseAgentListCliOutput / parseSkillsCliOutput — strip stdout before splitting, and decode JSON via a reviver that strips escapes from every string value.
3. SLUG_LINE_RE — widened to ^(\S+\/.*\S)\s*$ so custom/Model Two is recognised as a slug (fixesfix(opencode): parse model IDs containing spaces #7834).
4. loadInventoryFromCli error — when models exits non-zero, include truncated stderr/stdout in the OpenCodeRuntimeError detail instead of just the exit code, so Failed to execute OpenCode CLI health check: ... becomes diagnosable.

Verification

• Manual parser checks:
  • custom/Model Two now parses (old regex: fail, new: pass)
  • OSC ESC ]0;D:\test: ready BEL + CSI \u001b[31m stripped correctly
  • End-to-end parseModelsCliOutput with spaced IDs returns both models
• opencode models --verbose on opencode 1.18.23 (scoop) still exits 0; the improved error path now surfaces stderr if a future failure occurs.
• The fix matches the diffs of fix(opencode): parse model IDs containing spaces #7834 and fix(server): strip terminal escape sequences from opencode CLI output #7988 and passes their added tests conceptually.

Built with muse-spark-1.2-contributor via Opencode.

---

Note

Medium Risk
Parsing behavior changes for OpenCode CLI output (broader slug matching and escape stripping); incorrect stripping could drop valid data, but scope is limited to CLI integration paths.

Overview
Makes OpenCode CLI inventory and server startup parsing tolerate terminal escape noise and model IDs with spaces, and improves failures when opencode models exits non‑zero.

Adds stripTerminalEscapes in @t3tools/shared (OSC + CSI removal) and applies it in opencodeRuntime: server URL detection, models/agents/skills stdout line splitting, and a JSON.parse reviver so decoded strings never keep control bytes. SLUG_LINE_RE is relaxed so headers like custom/Kimi K3 count as provider/model slugs instead of being dropped.

parseSkillsCliOutput now strips the full stdout, parses JSON with the reviver, then validates with the existing skill schema (instead of decoding the raw CLI string as JSON in one step).

When loadInventoryFromCli sees a non‑zero models exit code, the error detail reports stderr/stdout lengths and attaches full stdout/stderr on cause for debugging.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Fix OpenCode health check parsing for terminal escapes and spaced model IDs

• Adds stripTerminalEscapes helper in packages/shared/src/stripTerminalEscapes.ts and exports it via @t3tools/shared/stripTerminalEscapes; it removes OSC/CSI control sequences from text.
• OpenCode CLI parsers in apps/server/src/provider/opencodeRuntime.ts (parseServerUrlFromOutput, parseModelsCliOutput, parseAgentListCliOutput, parseSkillsCliOutput) now sanitize stdout with stripTerminalEscapes, and JSON parsing uses a new reviver (stripTerminalEscapesJsonReviver) to strip escapes from embedded string fields at any depth.
• Loosens SLUG_LINE_RE from ^(\S+/\S+)\s*$ to ^(\S+/.\*\S)\s*$ so model IDs containing spaces after the slash still match.
• Reworks parseSkillsCliOutput to JSON-parse first (with the reviver) then validate via decodeOpenCodeSkillsArrayExit, which now expects a parsed value instead of a JSON string.
• Enriches loadInventoryFromCli non-zero-exit errors with stdout/stderr trimmed lengths in the detail and full stdout/stderr plus exitCode in a cause object.
• Behavioral Change: decodeOpenCodeSkillsArrayExit now takes a parsed value, so callers passing a raw JSON string will break; the only in-tree caller (parseSkillsCliOutput) is updated.

^{Macroscope summarized 2e6af7a.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 0e886f83-a5e4-48b5-95e0-5d6f6b4de33b

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

One finding: raw CLI stdout/stderr is copied into a caller-visible error detail. Error attributes should stay bounded and normalized, with the raw value preserved only as cause.

^{Posted via Macroscope — Effect Service Conventions}

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — This focused production fix updates OpenCode health-check parsing and terminal-output handling across existing server paths. Its failure-diagnostic change retains CLI output internally but does not expose it through the user-facing health-check detail, leaving a substantive behavior gap unresolved.

^{You can add or adjust custom eligibility rules. Learn more.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

Want fixes drafted automatically? Bugbot Autofix can create code changes for findings. A team admin can enable Autofix in the Cursor dashboard.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Configure here.}

[cursor]

CLI failure output never reaches users

Medium Severity

Non-zero opencode models now stores stdout and stderr on OpenCodeRuntimeError.cause and only puts character counts in detail. openCodeRuntimeErrorDetail and the health-check path read detail only, so Failed to execute OpenCode CLI health check still omits the CLI output needed to diagnose the failure.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Configure here.}

[t3dotgg]

Note

🤖 GPT-5.6 Sol responding on behalf of Theo

We're closing this PR as we clean up the T3 Code backlog. Thank you for taking the time to put this together.

We are keeping #7755 for terminal escape handling and #7834 for model IDs with spaces. This combined health-check branch does not need to remain open beside those focused fixes.

If you believe we closed this in error, please reopen the PR and leave a comment explaining what we missed.

[Mostafa-Ben-Git]

Closes#8386

Problem

opencode <= 1.18 emits an OSC title sequence (ESC ]0;<cwd>: ready BEL) on stdout even when piped, and some custom providers use model IDs with spaces (e.g. Kimi K3). On Windows this caused:

• parseServerUrlFromOutput to miss the opencode server listening line,
• parseModelsCliOutput to drop the model with a spaced ID and its predecessor (regex ^(\S+\/\S+)\s*$ doesn't allow spaces after the slash),
• parseAgentListCliOutput / parseSkillsCliOutput to retain raw control bytes inside decoded strings,
• and loadInventoryFromCli to surface OpenCode models command exited with code 1. with no stderr/stdout, making the failure unactionable.

All three parser bugs have open PRs (#7834, #7988, #8302, #7755) but were still present on main.

Fix

Ported from those PRs and combined:

1. New helperpackages/shared/src/stripTerminalEscapes.ts — removes OSC (terminated by BEL or ST) and CSI sequences.
2. parseServerUrlFromOutput / parseModelsCliOutput / parseAgentListCliOutput / parseSkillsCliOutput — strip stdout before splitting, and decode JSON via a reviver that strips escapes from every string value.
3. SLUG_LINE_RE — widened to ^(\S+\/.*\S)\s*$ so custom/Model Two is recognised as a slug (fixesfix(opencode): parse model IDs containing spaces #7834).
4. loadInventoryFromCli error — when models exits non-zero, include truncated stderr/stdout in the OpenCodeRuntimeError detail instead of just the exit code, so Failed to execute OpenCode CLI health check: ... becomes diagnosable.

Verification

• Manual parser checks:
  • custom/Model Two now parses (old regex: fail, new: pass)
  • OSC ESC ]0;D:\test: ready BEL + CSI \u001b[31m stripped correctly
  • End-to-end parseModelsCliOutput with spaced IDs returns both models
• opencode models --verbose on opencode 1.18.23 (scoop) still exits 0; the improved error path now surfaces stderr if a future failure occurs.
• The fix matches the diffs of fix(opencode): parse model IDs containing spaces #7834 and fix(server): strip terminal escape sequences from opencode CLI output #7988 and passes their added tests conceptually.

Built with muse-spark-1.2-contributor via Opencode.

---

Note

Medium Risk
Parsing behavior changes for OpenCode CLI output (broader slug matching and escape stripping); incorrect stripping could drop valid data, but scope is limited to CLI integration paths.

Overview
Makes OpenCode CLI inventory and server startup parsing tolerate terminal escape noise and model IDs with spaces, and improves failures when opencode models exits non‑zero.

Adds stripTerminalEscapes in @t3tools/shared (OSC + CSI removal) and applies it in opencodeRuntime: server URL detection, models/agents/skills stdout line splitting, and a JSON.parse reviver so decoded strings never keep control bytes. SLUG_LINE_RE is relaxed so headers like custom/Kimi K3 count as provider/model slugs instead of being dropped.

parseSkillsCliOutput now strips the full stdout, parses JSON with the reviver, then validates with the existing skill schema (instead of decoding the raw CLI string as JSON in one step).

When loadInventoryFromCli sees a non‑zero models exit code, the error detail reports stderr/stdout lengths and attaches full stdout/stderr on cause for debugging.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Fix OpenCode health check parsing for terminal escapes and spaced model IDs

• Adds stripTerminalEscapes helper in packages/shared/src/stripTerminalEscapes.ts and exports it via @t3tools/shared/stripTerminalEscapes; it removes OSC/CSI control sequences from text.
• OpenCode CLI parsers in apps/server/src/provider/opencodeRuntime.ts (parseServerUrlFromOutput, parseModelsCliOutput, parseAgentListCliOutput, parseSkillsCliOutput) now sanitize stdout with stripTerminalEscapes, and JSON parsing uses a new reviver (stripTerminalEscapesJsonReviver) to strip escapes from embedded string fields at any depth.
• Loosens SLUG_LINE_RE from ^(\S+/\S+)\s*$ to ^(\S+/.\*\S)\s*$ so model IDs containing spaces after the slash still match.
• Reworks parseSkillsCliOutput to JSON-parse first (with the reviver) then validate via decodeOpenCodeSkillsArrayExit, which now expects a parsed value instead of a JSON string.
• Enriches loadInventoryFromCli non-zero-exit errors with stdout/stderr trimmed lengths in the detail and full stdout/stderr plus exitCode in a cause object.
• Behavioral Change: decodeOpenCodeSkillsArrayExit now takes a parsed value, so callers passing a raw JSON string will break; the only in-tree caller (parseSkillsCliOutput) is updated.

^{Macroscope summarized 2e6af7a.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 0e886f83-a5e4-48b5-95e0-5d6f6b4de33b

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

One finding: raw CLI stdout/stderr is copied into a caller-visible error detail. Error attributes should stay bounded and normalized, with the raw value preserved only as cause.

^{Posted via Macroscope — Effect Service Conventions}

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — This focused production fix updates OpenCode health-check parsing and terminal-output handling across existing server paths. Its failure-diagnostic change retains CLI output internally but does not expose it through the user-facing health-check detail, leaving a substantive behavior gap unresolved.

^{You can add or adjust custom eligibility rules. Learn more.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

Want fixes drafted automatically? Bugbot Autofix can create code changes for findings. A team admin can enable Autofix in the Cursor dashboard.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Configure here.}

[cursor]

CLI failure output never reaches users

Medium Severity

Non-zero opencode models now stores stdout and stderr on OpenCodeRuntimeError.cause and only puts character counts in detail. openCodeRuntimeErrorDetail and the health-check path read detail only, so Failed to execute OpenCode CLI health check still omits the CLI output needed to diagnose the failure.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Configure here.}

[t3dotgg]

Note

🤖 GPT-5.6 Sol responding on behalf of Theo

We're closing this PR as we clean up the T3 Code backlog. Thank you for taking the time to put this together.

We are keeping #7755 for terminal escape handling and #7834 for model IDs with spaces. This combined health-check branch does not need to remain open beside those focused fixes.

If you believe we closed this in error, please reopen the PR and leave a comment explaining what we missed.

[Mostafa-Ben-Git]

Closes#8386

Problem

opencode <= 1.18 emits an OSC title sequence (ESC ]0;<cwd>: ready BEL) on stdout even when piped, and some custom providers use model IDs with spaces (e.g. Kimi K3). On Windows this caused:

• parseServerUrlFromOutput to miss the opencode server listening line,
• parseModelsCliOutput to drop the model with a spaced ID and its predecessor (regex ^(\S+\/\S+)\s*$ doesn't allow spaces after the slash),
• parseAgentListCliOutput / parseSkillsCliOutput to retain raw control bytes inside decoded strings,
• and loadInventoryFromCli to surface OpenCode models command exited with code 1. with no stderr/stdout, making the failure unactionable.

All three parser bugs have open PRs (#7834, #7988, #8302, #7755) but were still present on main.

Fix

Ported from those PRs and combined:

1. New helperpackages/shared/src/stripTerminalEscapes.ts — removes OSC (terminated by BEL or ST) and CSI sequences.
2. parseServerUrlFromOutput / parseModelsCliOutput / parseAgentListCliOutput / parseSkillsCliOutput — strip stdout before splitting, and decode JSON via a reviver that strips escapes from every string value.
3. SLUG_LINE_RE — widened to ^(\S+\/.*\S)\s*$ so custom/Model Two is recognised as a slug (fixesfix(opencode): parse model IDs containing spaces #7834).
4. loadInventoryFromCli error — when models exits non-zero, include truncated stderr/stdout in the OpenCodeRuntimeError detail instead of just the exit code, so Failed to execute OpenCode CLI health check: ... becomes diagnosable.

Verification

• Manual parser checks:
  • custom/Model Two now parses (old regex: fail, new: pass)
  • OSC ESC ]0;D:\test: ready BEL + CSI \u001b[31m stripped correctly
  • End-to-end parseModelsCliOutput with spaced IDs returns both models
• opencode models --verbose on opencode 1.18.23 (scoop) still exits 0; the improved error path now surfaces stderr if a future failure occurs.
• The fix matches the diffs of fix(opencode): parse model IDs containing spaces #7834 and fix(server): strip terminal escape sequences from opencode CLI output #7988 and passes their added tests conceptually.

Built with muse-spark-1.2-contributor via Opencode.

---

Note

Medium Risk
Parsing behavior changes for OpenCode CLI output (broader slug matching and escape stripping); incorrect stripping could drop valid data, but scope is limited to CLI integration paths.

Overview
Makes OpenCode CLI inventory and server startup parsing tolerate terminal escape noise and model IDs with spaces, and improves failures when opencode models exits non‑zero.

Adds stripTerminalEscapes in @t3tools/shared (OSC + CSI removal) and applies it in opencodeRuntime: server URL detection, models/agents/skills stdout line splitting, and a JSON.parse reviver so decoded strings never keep control bytes. SLUG_LINE_RE is relaxed so headers like custom/Kimi K3 count as provider/model slugs instead of being dropped.

parseSkillsCliOutput now strips the full stdout, parses JSON with the reviver, then validates with the existing skill schema (instead of decoding the raw CLI string as JSON in one step).

When loadInventoryFromCli sees a non‑zero models exit code, the error detail reports stderr/stdout lengths and attaches full stdout/stderr on cause for debugging.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Fix OpenCode health check parsing for terminal escapes and spaced model IDs

• Adds stripTerminalEscapes helper in packages/shared/src/stripTerminalEscapes.ts and exports it via @t3tools/shared/stripTerminalEscapes; it removes OSC/CSI control sequences from text.
• OpenCode CLI parsers in apps/server/src/provider/opencodeRuntime.ts (parseServerUrlFromOutput, parseModelsCliOutput, parseAgentListCliOutput, parseSkillsCliOutput) now sanitize stdout with stripTerminalEscapes, and JSON parsing uses a new reviver (stripTerminalEscapesJsonReviver) to strip escapes from embedded string fields at any depth.
• Loosens SLUG_LINE_RE from ^(\S+/\S+)\s*$ to ^(\S+/.\*\S)\s*$ so model IDs containing spaces after the slash still match.
• Reworks parseSkillsCliOutput to JSON-parse first (with the reviver) then validate via decodeOpenCodeSkillsArrayExit, which now expects a parsed value instead of a JSON string.
• Enriches loadInventoryFromCli non-zero-exit errors with stdout/stderr trimmed lengths in the detail and full stdout/stderr plus exitCode in a cause object.
• Behavioral Change: decodeOpenCodeSkillsArrayExit now takes a parsed value, so callers passing a raw JSON string will break; the only in-tree caller (parseSkillsCliOutput) is updated.

^{Macroscope summarized 2e6af7a.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 0e886f83-a5e4-48b5-95e0-5d6f6b4de33b

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

One finding: raw CLI stdout/stderr is copied into a caller-visible error detail. Error attributes should stay bounded and normalized, with the raw value preserved only as cause.

^{Posted via Macroscope — Effect Service Conventions}

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — This focused production fix updates OpenCode health-check parsing and terminal-output handling across existing server paths. Its failure-diagnostic change retains CLI output internally but does not expose it through the user-facing health-check detail, leaving a substantive behavior gap unresolved.

^{You can add or adjust custom eligibility rules. Learn more.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

Want fixes drafted automatically? Bugbot Autofix can create code changes for findings. A team admin can enable Autofix in the Cursor dashboard.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Configure here.}

[cursor]

CLI failure output never reaches users

Medium Severity

Non-zero opencode models now stores stdout and stderr on OpenCodeRuntimeError.cause and only puts character counts in detail. openCodeRuntimeErrorDetail and the health-check path read detail only, so Failed to execute OpenCode CLI health check still omits the CLI output needed to diagnose the failure.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Configure here.}

[t3dotgg]

Note

🤖 GPT-5.6 Sol responding on behalf of Theo

We're closing this PR as we clean up the T3 Code backlog. Thank you for taking the time to put this together.

We are keeping #7755 for terminal escape handling and #7834 for model IDs with spaces. This combined health-check branch does not need to remain open beside those focused fixes.

If you believe we closed this in error, please reopen the PR and leave a comment explaining what we missed.

[Mostafa-Ben-Git]

Closes#8386

Problem

opencode <= 1.18 emits an OSC title sequence (ESC ]0;<cwd>: ready BEL) on stdout even when piped, and some custom providers use model IDs with spaces (e.g. Kimi K3). On Windows this caused:

• parseServerUrlFromOutput to miss the opencode server listening line,
• parseModelsCliOutput to drop the model with a spaced ID and its predecessor (regex ^(\S+\/\S+)\s*$ doesn't allow spaces after the slash),
• parseAgentListCliOutput / parseSkillsCliOutput to retain raw control bytes inside decoded strings,
• and loadInventoryFromCli to surface OpenCode models command exited with code 1. with no stderr/stdout, making the failure unactionable.

All three parser bugs have open PRs (#7834, #7988, #8302, #7755) but were still present on main.

Fix

Ported from those PRs and combined:

1. New helperpackages/shared/src/stripTerminalEscapes.ts — removes OSC (terminated by BEL or ST) and CSI sequences.
2. parseServerUrlFromOutput / parseModelsCliOutput / parseAgentListCliOutput / parseSkillsCliOutput — strip stdout before splitting, and decode JSON via a reviver that strips escapes from every string value.
3. SLUG_LINE_RE — widened to ^(\S+\/.*\S)\s*$ so custom/Model Two is recognised as a slug (fixesfix(opencode): parse model IDs containing spaces #7834).
4. loadInventoryFromCli error — when models exits non-zero, include truncated stderr/stdout in the OpenCodeRuntimeError detail instead of just the exit code, so Failed to execute OpenCode CLI health check: ... becomes diagnosable.

Verification

• Manual parser checks:
  • custom/Model Two now parses (old regex: fail, new: pass)
  • OSC ESC ]0;D:\test: ready BEL + CSI \u001b[31m stripped correctly
  • End-to-end parseModelsCliOutput with spaced IDs returns both models
• opencode models --verbose on opencode 1.18.23 (scoop) still exits 0; the improved error path now surfaces stderr if a future failure occurs.
• The fix matches the diffs of fix(opencode): parse model IDs containing spaces #7834 and fix(server): strip terminal escape sequences from opencode CLI output #7988 and passes their added tests conceptually.

Built with muse-spark-1.2-contributor via Opencode.

---

Note

Medium Risk
Parsing behavior changes for OpenCode CLI output (broader slug matching and escape stripping); incorrect stripping could drop valid data, but scope is limited to CLI integration paths.

Overview
Makes OpenCode CLI inventory and server startup parsing tolerate terminal escape noise and model IDs with spaces, and improves failures when opencode models exits non‑zero.

Adds stripTerminalEscapes in @t3tools/shared (OSC + CSI removal) and applies it in opencodeRuntime: server URL detection, models/agents/skills stdout line splitting, and a JSON.parse reviver so decoded strings never keep control bytes. SLUG_LINE_RE is relaxed so headers like custom/Kimi K3 count as provider/model slugs instead of being dropped.

parseSkillsCliOutput now strips the full stdout, parses JSON with the reviver, then validates with the existing skill schema (instead of decoding the raw CLI string as JSON in one step).

When loadInventoryFromCli sees a non‑zero models exit code, the error detail reports stderr/stdout lengths and attaches full stdout/stderr on cause for debugging.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Fix OpenCode health check parsing for terminal escapes and spaced model IDs

• Adds stripTerminalEscapes helper in packages/shared/src/stripTerminalEscapes.ts and exports it via @t3tools/shared/stripTerminalEscapes; it removes OSC/CSI control sequences from text.
• OpenCode CLI parsers in apps/server/src/provider/opencodeRuntime.ts (parseServerUrlFromOutput, parseModelsCliOutput, parseAgentListCliOutput, parseSkillsCliOutput) now sanitize stdout with stripTerminalEscapes, and JSON parsing uses a new reviver (stripTerminalEscapesJsonReviver) to strip escapes from embedded string fields at any depth.
• Loosens SLUG_LINE_RE from ^(\S+/\S+)\s*$ to ^(\S+/.\*\S)\s*$ so model IDs containing spaces after the slash still match.
• Reworks parseSkillsCliOutput to JSON-parse first (with the reviver) then validate via decodeOpenCodeSkillsArrayExit, which now expects a parsed value instead of a JSON string.
• Enriches loadInventoryFromCli non-zero-exit errors with stdout/stderr trimmed lengths in the detail and full stdout/stderr plus exitCode in a cause object.
• Behavioral Change: decodeOpenCodeSkillsArrayExit now takes a parsed value, so callers passing a raw JSON string will break; the only in-tree caller (parseSkillsCliOutput) is updated.

^{Macroscope summarized 2e6af7a.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 0e886f83-a5e4-48b5-95e0-5d6f6b4de33b

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

One finding: raw CLI stdout/stderr is copied into a caller-visible error detail. Error attributes should stay bounded and normalized, with the raw value preserved only as cause.

^{Posted via Macroscope — Effect Service Conventions}

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — This focused production fix updates OpenCode health-check parsing and terminal-output handling across existing server paths. Its failure-diagnostic change retains CLI output internally but does not expose it through the user-facing health-check detail, leaving a substantive behavior gap unresolved.

^{You can add or adjust custom eligibility rules. Learn more.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

Want fixes drafted automatically? Bugbot Autofix can create code changes for findings. A team admin can enable Autofix in the Cursor dashboard.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Configure here.}

[cursor]

CLI failure output never reaches users

Medium Severity

Non-zero opencode models now stores stdout and stderr on OpenCodeRuntimeError.cause and only puts character counts in detail. openCodeRuntimeErrorDetail and the health-check path read detail only, so Failed to execute OpenCode CLI health check still omits the CLI output needed to diagnose the failure.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Configure here.}

[t3dotgg]

Note

🤖 GPT-5.6 Sol responding on behalf of Theo

We're closing this PR as we clean up the T3 Code backlog. Thank you for taking the time to put this together.

We are keeping #7755 for terminal escape handling and #7834 for model IDs with spaces. This combined health-check branch does not need to remain open beside those focused fixes.

If you believe we closed this in error, please reopen the PR and leave a comment explaining what we missed.

[Mostafa-Ben-Git]

Closes#8386

Problem

opencode <= 1.18 emits an OSC title sequence (ESC ]0;<cwd>: ready BEL) on stdout even when piped, and some custom providers use model IDs with spaces (e.g. Kimi K3). On Windows this caused:

• parseServerUrlFromOutput to miss the opencode server listening line,
• parseModelsCliOutput to drop the model with a spaced ID and its predecessor (regex ^(\S+\/\S+)\s*$ doesn't allow spaces after the slash),
• parseAgentListCliOutput / parseSkillsCliOutput to retain raw control bytes inside decoded strings,
• and loadInventoryFromCli to surface OpenCode models command exited with code 1. with no stderr/stdout, making the failure unactionable.

All three parser bugs have open PRs (#7834, #7988, #8302, #7755) but were still present on main.

Fix

Ported from those PRs and combined:

1. New helperpackages/shared/src/stripTerminalEscapes.ts — removes OSC (terminated by BEL or ST) and CSI sequences.
2. parseServerUrlFromOutput / parseModelsCliOutput / parseAgentListCliOutput / parseSkillsCliOutput — strip stdout before splitting, and decode JSON via a reviver that strips escapes from every string value.
3. SLUG_LINE_RE — widened to ^(\S+\/.*\S)\s*$ so custom/Model Two is recognised as a slug (fixesfix(opencode): parse model IDs containing spaces #7834).
4. loadInventoryFromCli error — when models exits non-zero, include truncated stderr/stdout in the OpenCodeRuntimeError detail instead of just the exit code, so Failed to execute OpenCode CLI health check: ... becomes diagnosable.

Verification

• Manual parser checks:
  • custom/Model Two now parses (old regex: fail, new: pass)
  • OSC ESC ]0;D:\test: ready BEL + CSI \u001b[31m stripped correctly
  • End-to-end parseModelsCliOutput with spaced IDs returns both models
• opencode models --verbose on opencode 1.18.23 (scoop) still exits 0; the improved error path now surfaces stderr if a future failure occurs.
• The fix matches the diffs of fix(opencode): parse model IDs containing spaces #7834 and fix(server): strip terminal escape sequences from opencode CLI output #7988 and passes their added tests conceptually.

Built with muse-spark-1.2-contributor via Opencode.

---

Note

Medium Risk
Parsing behavior changes for OpenCode CLI output (broader slug matching and escape stripping); incorrect stripping could drop valid data, but scope is limited to CLI integration paths.

Overview
Makes OpenCode CLI inventory and server startup parsing tolerate terminal escape noise and model IDs with spaces, and improves failures when opencode models exits non‑zero.

Adds stripTerminalEscapes in @t3tools/shared (OSC + CSI removal) and applies it in opencodeRuntime: server URL detection, models/agents/skills stdout line splitting, and a JSON.parse reviver so decoded strings never keep control bytes. SLUG_LINE_RE is relaxed so headers like custom/Kimi K3 count as provider/model slugs instead of being dropped.

parseSkillsCliOutput now strips the full stdout, parses JSON with the reviver, then validates with the existing skill schema (instead of decoding the raw CLI string as JSON in one step).

When loadInventoryFromCli sees a non‑zero models exit code, the error detail reports stderr/stdout lengths and attaches full stdout/stderr on cause for debugging.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Fix OpenCode health check parsing for terminal escapes and spaced model IDs

• Adds stripTerminalEscapes helper in packages/shared/src/stripTerminalEscapes.ts and exports it via @t3tools/shared/stripTerminalEscapes; it removes OSC/CSI control sequences from text.
• OpenCode CLI parsers in apps/server/src/provider/opencodeRuntime.ts (parseServerUrlFromOutput, parseModelsCliOutput, parseAgentListCliOutput, parseSkillsCliOutput) now sanitize stdout with stripTerminalEscapes, and JSON parsing uses a new reviver (stripTerminalEscapesJsonReviver) to strip escapes from embedded string fields at any depth.
• Loosens SLUG_LINE_RE from ^(\S+/\S+)\s*$ to ^(\S+/.\*\S)\s*$ so model IDs containing spaces after the slash still match.
• Reworks parseSkillsCliOutput to JSON-parse first (with the reviver) then validate via decodeOpenCodeSkillsArrayExit, which now expects a parsed value instead of a JSON string.
• Enriches loadInventoryFromCli non-zero-exit errors with stdout/stderr trimmed lengths in the detail and full stdout/stderr plus exitCode in a cause object.
• Behavioral Change: decodeOpenCodeSkillsArrayExit now takes a parsed value, so callers passing a raw JSON string will break; the only in-tree caller (parseSkillsCliOutput) is updated.

^{Macroscope summarized 2e6af7a.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 0e886f83-a5e4-48b5-95e0-5d6f6b4de33b

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

One finding: raw CLI stdout/stderr is copied into a caller-visible error detail. Error attributes should stay bounded and normalized, with the raw value preserved only as cause.

^{Posted via Macroscope — Effect Service Conventions}

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — This focused production fix updates OpenCode health-check parsing and terminal-output handling across existing server paths. Its failure-diagnostic change retains CLI output internally but does not expose it through the user-facing health-check detail, leaving a substantive behavior gap unresolved.

^{You can add or adjust custom eligibility rules. Learn more.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

Want fixes drafted automatically? Bugbot Autofix can create code changes for findings. A team admin can enable Autofix in the Cursor dashboard.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Configure here.}

[cursor]

CLI failure output never reaches users

Medium Severity

Non-zero opencode models now stores stdout and stderr on OpenCodeRuntimeError.cause and only puts character counts in detail. openCodeRuntimeErrorDetail and the health-check path read detail only, so Failed to execute OpenCode CLI health check still omits the CLI output needed to diagnose the failure.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Configure here.}

[t3dotgg]

Note

🤖 GPT-5.6 Sol responding on behalf of Theo

We're closing this PR as we clean up the T3 Code backlog. Thank you for taking the time to put this together.

We are keeping #7755 for terminal escape handling and #7834 for model IDs with spaces. This combined health-check branch does not need to remain open beside those focused fixes.

If you believe we closed this in error, please reopen the PR and leave a comment explaining what we missed.

[Mostafa-Ben-Git]

Closes#8386

Problem

opencode <= 1.18 emits an OSC title sequence (ESC ]0;<cwd>: ready BEL) on stdout even when piped, and some custom providers use model IDs with spaces (e.g. Kimi K3). On Windows this caused:

• parseServerUrlFromOutput to miss the opencode server listening line,
• parseModelsCliOutput to drop the model with a spaced ID and its predecessor (regex ^(\S+\/\S+)\s*$ doesn't allow spaces after the slash),
• parseAgentListCliOutput / parseSkillsCliOutput to retain raw control bytes inside decoded strings,
• and loadInventoryFromCli to surface OpenCode models command exited with code 1. with no stderr/stdout, making the failure unactionable.

All three parser bugs have open PRs (#7834, #7988, #8302, #7755) but were still present on main.

Fix

Ported from those PRs and combined:

1. New helperpackages/shared/src/stripTerminalEscapes.ts — removes OSC (terminated by BEL or ST) and CSI sequences.
2. parseServerUrlFromOutput / parseModelsCliOutput / parseAgentListCliOutput / parseSkillsCliOutput — strip stdout before splitting, and decode JSON via a reviver that strips escapes from every string value.
3. SLUG_LINE_RE — widened to ^(\S+\/.*\S)\s*$ so custom/Model Two is recognised as a slug (fixesfix(opencode): parse model IDs containing spaces #7834).
4. loadInventoryFromCli error — when models exits non-zero, include truncated stderr/stdout in the OpenCodeRuntimeError detail instead of just the exit code, so Failed to execute OpenCode CLI health check: ... becomes diagnosable.

Verification

• Manual parser checks:
  • custom/Model Two now parses (old regex: fail, new: pass)
  • OSC ESC ]0;D:\test: ready BEL + CSI \u001b[31m stripped correctly
  • End-to-end parseModelsCliOutput with spaced IDs returns both models
• opencode models --verbose on opencode 1.18.23 (scoop) still exits 0; the improved error path now surfaces stderr if a future failure occurs.
• The fix matches the diffs of fix(opencode): parse model IDs containing spaces #7834 and fix(server): strip terminal escape sequences from opencode CLI output #7988 and passes their added tests conceptually.

Built with muse-spark-1.2-contributor via Opencode.

---

Note

Medium Risk
Parsing behavior changes for OpenCode CLI output (broader slug matching and escape stripping); incorrect stripping could drop valid data, but scope is limited to CLI integration paths.

Overview
Makes OpenCode CLI inventory and server startup parsing tolerate terminal escape noise and model IDs with spaces, and improves failures when opencode models exits non‑zero.

Adds stripTerminalEscapes in @t3tools/shared (OSC + CSI removal) and applies it in opencodeRuntime: server URL detection, models/agents/skills stdout line splitting, and a JSON.parse reviver so decoded strings never keep control bytes. SLUG_LINE_RE is relaxed so headers like custom/Kimi K3 count as provider/model slugs instead of being dropped.

parseSkillsCliOutput now strips the full stdout, parses JSON with the reviver, then validates with the existing skill schema (instead of decoding the raw CLI string as JSON in one step).

When loadInventoryFromCli sees a non‑zero models exit code, the error detail reports stderr/stdout lengths and attaches full stdout/stderr on cause for debugging.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Fix OpenCode health check parsing for terminal escapes and spaced model IDs

• Adds stripTerminalEscapes helper in packages/shared/src/stripTerminalEscapes.ts and exports it via @t3tools/shared/stripTerminalEscapes; it removes OSC/CSI control sequences from text.
• OpenCode CLI parsers in apps/server/src/provider/opencodeRuntime.ts (parseServerUrlFromOutput, parseModelsCliOutput, parseAgentListCliOutput, parseSkillsCliOutput) now sanitize stdout with stripTerminalEscapes, and JSON parsing uses a new reviver (stripTerminalEscapesJsonReviver) to strip escapes from embedded string fields at any depth.
• Loosens SLUG_LINE_RE from ^(\S+/\S+)\s*$ to ^(\S+/.\*\S)\s*$ so model IDs containing spaces after the slash still match.
• Reworks parseSkillsCliOutput to JSON-parse first (with the reviver) then validate via decodeOpenCodeSkillsArrayExit, which now expects a parsed value instead of a JSON string.
• Enriches loadInventoryFromCli non-zero-exit errors with stdout/stderr trimmed lengths in the detail and full stdout/stderr plus exitCode in a cause object.
• Behavioral Change: decodeOpenCodeSkillsArrayExit now takes a parsed value, so callers passing a raw JSON string will break; the only in-tree caller (parseSkillsCliOutput) is updated.

^{Macroscope summarized 2e6af7a.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 0e886f83-a5e4-48b5-95e0-5d6f6b4de33b

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

One finding: raw CLI stdout/stderr is copied into a caller-visible error detail. Error attributes should stay bounded and normalized, with the raw value preserved only as cause.

^{Posted via Macroscope — Effect Service Conventions}

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — This focused production fix updates OpenCode health-check parsing and terminal-output handling across existing server paths. Its failure-diagnostic change retains CLI output internally but does not expose it through the user-facing health-check detail, leaving a substantive behavior gap unresolved.

^{You can add or adjust custom eligibility rules. Learn more.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

Want fixes drafted automatically? Bugbot Autofix can create code changes for findings. A team admin can enable Autofix in the Cursor dashboard.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Configure here.}

[cursor]

CLI failure output never reaches users

Medium Severity

Non-zero opencode models now stores stdout and stderr on OpenCodeRuntimeError.cause and only puts character counts in detail. openCodeRuntimeErrorDetail and the health-check path read detail only, so Failed to execute OpenCode CLI health check still omits the CLI output needed to diagnose the failure.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Configure here.}

[t3dotgg]

Note

🤖 GPT-5.6 Sol responding on behalf of Theo

We're closing this PR as we clean up the T3 Code backlog. Thank you for taking the time to put this together.

We are keeping #7755 for terminal escape handling and #7834 for model IDs with spaces. This combined health-check branch does not need to remain open beside those focused fixes.

If you believe we closed this in error, please reopen the PR and leave a comment explaining what we missed.

[Mostafa-Ben-Git]

Closes#8386

Problem

opencode <= 1.18 emits an OSC title sequence (ESC ]0;<cwd>: ready BEL) on stdout even when piped, and some custom providers use model IDs with spaces (e.g. Kimi K3). On Windows this caused:

• parseServerUrlFromOutput to miss the opencode server listening line,
• parseModelsCliOutput to drop the model with a spaced ID and its predecessor (regex ^(\S+\/\S+)\s*$ doesn't allow spaces after the slash),
• parseAgentListCliOutput / parseSkillsCliOutput to retain raw control bytes inside decoded strings,
• and loadInventoryFromCli to surface OpenCode models command exited with code 1. with no stderr/stdout, making the failure unactionable.

All three parser bugs have open PRs (#7834, #7988, #8302, #7755) but were still present on main.

Fix

Ported from those PRs and combined:

1. New helperpackages/shared/src/stripTerminalEscapes.ts — removes OSC (terminated by BEL or ST) and CSI sequences.
2. parseServerUrlFromOutput / parseModelsCliOutput / parseAgentListCliOutput / parseSkillsCliOutput — strip stdout before splitting, and decode JSON via a reviver that strips escapes from every string value.
3. SLUG_LINE_RE — widened to ^(\S+\/.*\S)\s*$ so custom/Model Two is recognised as a slug (fixesfix(opencode): parse model IDs containing spaces #7834).
4. loadInventoryFromCli error — when models exits non-zero, include truncated stderr/stdout in the OpenCodeRuntimeError detail instead of just the exit code, so Failed to execute OpenCode CLI health check: ... becomes diagnosable.

Verification

• Manual parser checks:
  • custom/Model Two now parses (old regex: fail, new: pass)
  • OSC ESC ]0;D:\test: ready BEL + CSI \u001b[31m stripped correctly
  • End-to-end parseModelsCliOutput with spaced IDs returns both models
• opencode models --verbose on opencode 1.18.23 (scoop) still exits 0; the improved error path now surfaces stderr if a future failure occurs.
• The fix matches the diffs of fix(opencode): parse model IDs containing spaces #7834 and fix(server): strip terminal escape sequences from opencode CLI output #7988 and passes their added tests conceptually.

Built with muse-spark-1.2-contributor via Opencode.

---

Note

Medium Risk
Parsing behavior changes for OpenCode CLI output (broader slug matching and escape stripping); incorrect stripping could drop valid data, but scope is limited to CLI integration paths.

Overview
Makes OpenCode CLI inventory and server startup parsing tolerate terminal escape noise and model IDs with spaces, and improves failures when opencode models exits non‑zero.

Adds stripTerminalEscapes in @t3tools/shared (OSC + CSI removal) and applies it in opencodeRuntime: server URL detection, models/agents/skills stdout line splitting, and a JSON.parse reviver so decoded strings never keep control bytes. SLUG_LINE_RE is relaxed so headers like custom/Kimi K3 count as provider/model slugs instead of being dropped.

parseSkillsCliOutput now strips the full stdout, parses JSON with the reviver, then validates with the existing skill schema (instead of decoding the raw CLI string as JSON in one step).

When loadInventoryFromCli sees a non‑zero models exit code, the error detail reports stderr/stdout lengths and attaches full stdout/stderr on cause for debugging.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Fix OpenCode health check parsing for terminal escapes and spaced model IDs

• Adds stripTerminalEscapes helper in packages/shared/src/stripTerminalEscapes.ts and exports it via @t3tools/shared/stripTerminalEscapes; it removes OSC/CSI control sequences from text.
• OpenCode CLI parsers in apps/server/src/provider/opencodeRuntime.ts (parseServerUrlFromOutput, parseModelsCliOutput, parseAgentListCliOutput, parseSkillsCliOutput) now sanitize stdout with stripTerminalEscapes, and JSON parsing uses a new reviver (stripTerminalEscapesJsonReviver) to strip escapes from embedded string fields at any depth.
• Loosens SLUG_LINE_RE from ^(\S+/\S+)\s*$ to ^(\S+/.\*\S)\s*$ so model IDs containing spaces after the slash still match.
• Reworks parseSkillsCliOutput to JSON-parse first (with the reviver) then validate via decodeOpenCodeSkillsArrayExit, which now expects a parsed value instead of a JSON string.
• Enriches loadInventoryFromCli non-zero-exit errors with stdout/stderr trimmed lengths in the detail and full stdout/stderr plus exitCode in a cause object.
• Behavioral Change: decodeOpenCodeSkillsArrayExit now takes a parsed value, so callers passing a raw JSON string will break; the only in-tree caller (parseSkillsCliOutput) is updated.

^{Macroscope summarized 2e6af7a.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 0e886f83-a5e4-48b5-95e0-5d6f6b4de33b

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

One finding: raw CLI stdout/stderr is copied into a caller-visible error detail. Error attributes should stay bounded and normalized, with the raw value preserved only as cause.

^{Posted via Macroscope — Effect Service Conventions}

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — This focused production fix updates OpenCode health-check parsing and terminal-output handling across existing server paths. Its failure-diagnostic change retains CLI output internally but does not expose it through the user-facing health-check detail, leaving a substantive behavior gap unresolved.

^{You can add or adjust custom eligibility rules. Learn more.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

Want fixes drafted automatically? Bugbot Autofix can create code changes for findings. A team admin can enable Autofix in the Cursor dashboard.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Configure here.}

[cursor]

CLI failure output never reaches users

Medium Severity

Non-zero opencode models now stores stdout and stderr on OpenCodeRuntimeError.cause and only puts character counts in detail. openCodeRuntimeErrorDetail and the health-check path read detail only, so Failed to execute OpenCode CLI health check still omits the CLI output needed to diagnose the failure.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Configure here.}

[t3dotgg]

Note

🤖 GPT-5.6 Sol responding on behalf of Theo

We're closing this PR as we clean up the T3 Code backlog. Thank you for taking the time to put this together.

We are keeping #7755 for terminal escape handling and #7834 for model IDs with spaces. This combined health-check branch does not need to remain open beside those focused fixes.

If you believe we closed this in error, please reopen the PR and leave a comment explaining what we missed.

[Mostafa-Ben-Git]

Closes#8386

Problem

opencode <= 1.18 emits an OSC title sequence (ESC ]0;<cwd>: ready BEL) on stdout even when piped, and some custom providers use model IDs with spaces (e.g. Kimi K3). On Windows this caused:

• parseServerUrlFromOutput to miss the opencode server listening line,
• parseModelsCliOutput to drop the model with a spaced ID and its predecessor (regex ^(\S+\/\S+)\s*$ doesn't allow spaces after the slash),
• parseAgentListCliOutput / parseSkillsCliOutput to retain raw control bytes inside decoded strings,
• and loadInventoryFromCli to surface OpenCode models command exited with code 1. with no stderr/stdout, making the failure unactionable.

All three parser bugs have open PRs (#7834, #7988, #8302, #7755) but were still present on main.

Fix

Ported from those PRs and combined:

1. New helperpackages/shared/src/stripTerminalEscapes.ts — removes OSC (terminated by BEL or ST) and CSI sequences.
2. parseServerUrlFromOutput / parseModelsCliOutput / parseAgentListCliOutput / parseSkillsCliOutput — strip stdout before splitting, and decode JSON via a reviver that strips escapes from every string value.
3. SLUG_LINE_RE — widened to ^(\S+\/.*\S)\s*$ so custom/Model Two is recognised as a slug (fixesfix(opencode): parse model IDs containing spaces #7834).
4. loadInventoryFromCli error — when models exits non-zero, include truncated stderr/stdout in the OpenCodeRuntimeError detail instead of just the exit code, so Failed to execute OpenCode CLI health check: ... becomes diagnosable.

Verification

• Manual parser checks:
  • custom/Model Two now parses (old regex: fail, new: pass)
  • OSC ESC ]0;D:\test: ready BEL + CSI \u001b[31m stripped correctly
  • End-to-end parseModelsCliOutput with spaced IDs returns both models
• opencode models --verbose on opencode 1.18.23 (scoop) still exits 0; the improved error path now surfaces stderr if a future failure occurs.
• The fix matches the diffs of fix(opencode): parse model IDs containing spaces #7834 and fix(server): strip terminal escape sequences from opencode CLI output #7988 and passes their added tests conceptually.

Built with muse-spark-1.2-contributor via Opencode.

---

Note

Medium Risk
Parsing behavior changes for OpenCode CLI output (broader slug matching and escape stripping); incorrect stripping could drop valid data, but scope is limited to CLI integration paths.

Overview
Makes OpenCode CLI inventory and server startup parsing tolerate terminal escape noise and model IDs with spaces, and improves failures when opencode models exits non‑zero.

Adds stripTerminalEscapes in @t3tools/shared (OSC + CSI removal) and applies it in opencodeRuntime: server URL detection, models/agents/skills stdout line splitting, and a JSON.parse reviver so decoded strings never keep control bytes. SLUG_LINE_RE is relaxed so headers like custom/Kimi K3 count as provider/model slugs instead of being dropped.

parseSkillsCliOutput now strips the full stdout, parses JSON with the reviver, then validates with the existing skill schema (instead of decoding the raw CLI string as JSON in one step).

When loadInventoryFromCli sees a non‑zero models exit code, the error detail reports stderr/stdout lengths and attaches full stdout/stderr on cause for debugging.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Fix OpenCode health check parsing for terminal escapes and spaced model IDs

• Adds stripTerminalEscapes helper in packages/shared/src/stripTerminalEscapes.ts and exports it via @t3tools/shared/stripTerminalEscapes; it removes OSC/CSI control sequences from text.
• OpenCode CLI parsers in apps/server/src/provider/opencodeRuntime.ts (parseServerUrlFromOutput, parseModelsCliOutput, parseAgentListCliOutput, parseSkillsCliOutput) now sanitize stdout with stripTerminalEscapes, and JSON parsing uses a new reviver (stripTerminalEscapesJsonReviver) to strip escapes from embedded string fields at any depth.
• Loosens SLUG_LINE_RE from ^(\S+/\S+)\s*$ to ^(\S+/.\*\S)\s*$ so model IDs containing spaces after the slash still match.
• Reworks parseSkillsCliOutput to JSON-parse first (with the reviver) then validate via decodeOpenCodeSkillsArrayExit, which now expects a parsed value instead of a JSON string.
• Enriches loadInventoryFromCli non-zero-exit errors with stdout/stderr trimmed lengths in the detail and full stdout/stderr plus exitCode in a cause object.
• Behavioral Change: decodeOpenCodeSkillsArrayExit now takes a parsed value, so callers passing a raw JSON string will break; the only in-tree caller (parseSkillsCliOutput) is updated.

^{Macroscope summarized 2e6af7a.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 0e886f83-a5e4-48b5-95e0-5d6f6b4de33b

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

One finding: raw CLI stdout/stderr is copied into a caller-visible error detail. Error attributes should stay bounded and normalized, with the raw value preserved only as cause.

^{Posted via Macroscope — Effect Service Conventions}

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — This focused production fix updates OpenCode health-check parsing and terminal-output handling across existing server paths. Its failure-diagnostic change retains CLI output internally but does not expose it through the user-facing health-check detail, leaving a substantive behavior gap unresolved.

^{You can add or adjust custom eligibility rules. Learn more.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

Want fixes drafted automatically? Bugbot Autofix can create code changes for findings. A team admin can enable Autofix in the Cursor dashboard.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Configure here.}

[cursor]

CLI failure output never reaches users

Medium Severity

Non-zero opencode models now stores stdout and stderr on OpenCodeRuntimeError.cause and only puts character counts in detail. openCodeRuntimeErrorDetail and the health-check path read detail only, so Failed to execute OpenCode CLI health check still omits the CLI output needed to diagnose the failure.

^{Reviewed by Cursor Bugbot for commit 2e6af7a. Configure here.}

[t3dotgg]

Note

🤖 GPT-5.6 Sol responding on behalf of Theo

We're closing this PR as we clean up the T3 Code backlog. Thank you for taking the time to put this together.

We are keeping #7755 for terminal escape handling and #7834 for model IDs with spaces. This combined health-check branch does not need to remain open beside those focused fixes.

If you believe we closed this in error, please reopen the PR and leave a comment explaining what we missed.