Uh oh!
There was an error while loading. Please reload this page.
Automatic capture works out of the box on Codex CLI and Desktop - #55
Merged
Conversation
Live testing from a genuinely reverted, out-of-the-box Codex config (no prior trust state, no custom profiles) found that neither the workaround documented in PR #53/#54 nor issue #47's "not wired" conclusion holds up: - Codex CLI shows a native "Hooks need review" dialog on first use in a project with throughline installed - choosing "Trust all and continue" is the entire setup step. - Codex Desktop grants trust silently, no dialog at all. Both confirmed via real captured buffer content, not just config state. Replaces the "Advanced: enabling automatic capture (unsupported)" README section and the promo site's "no automatic capture - trust-gated" framing with the actual current behavior: full parity with Claude Code (skills + all 5 hooks), gated only by a one-time trust step that's usually invisible. .claude-plugin and .codex-plugin manifests are unchanged - this was always true of the shipped plugin; the prior docs just had it wrong. Closes#52. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013cG9PVNGundwPjNB4Yz5AS
/review-pr on PR #55 found: - CHANGELOG.md's v0.13.0 entry still asserted "skills only, not automatic capture" and "so the skills-only framing stands" for the currently-shipped release - a false statement about installed code, not just stale history. Added an [Unreleased] correction entry and bracketed corrections inline in the released 0.13.0 block (never rewriting released history silently). - The new PreCompact paragraph narrated the investigation ("hasn't been directly observed") rather than stating current behavior - violates the writing-style rule added to global CLAUDE.md this same session. Dropped the hedge; PreCompact installs and runs identically to the other four events already confirmed firing, so there's no basis to single it out. This also fixes the matrix/prose inconsistency the review flagged (Compaction survival: yes in the table, hedged in prose). - "isn't asked again for this project or any other" overstated permanence - the dialog's own text says "new or changed," meaning a throughline release that touches hook scripts re-triggers it. Corrected. - Neither the README's Codex section nor the promo site's Codex card stated the jq requirement, despite both matrices now claiming it - added Requirements notes matching the Claude Code sections' existing pattern. - Trimmed an overlong layout-tree comment that broke horizontal alignment. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013cG9PVNGundwPjNB4Yz5AS
Uh oh!
There was an error while loading. Please reload this page.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for freeto join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Corrects the README and promo site to reflect actual current Codex behavior, confirmed live from a genuinely out-of-the-box config (no prior trust state, no custom profiles):
Both confirmed via real captured buffer content in fresh test projects, not just config inspection.
Replaces the "Advanced: enabling automatic capture (unsupported)" README section (added in #53/#54, since obsolete) and the promo site's "no automatic capture - trust-gated" framing with the actual current behavior: full parity with Claude Code, gated only by a usually-invisible one-time trust step.
No code changes —
.claude-plugin/.codex-pluginmanifests are untouched. This was always true of the shipped plugin; the prior docs had it wrong.Closes#52.
Test plan
/local-ci— 167/167 pass, manifest checks passcodex exec/interactive CLI session and fresh Desktop session both captured real buffer content from a reverted, untouched Codex config🤖 Generated with Claude Code
https://claude.ai/code/session_013cG9PVNGundwPjNB4Yz5AS