Sync vendored skills with skills main (f8959e14 / #63) - #23
Conversation
Vendor-sync plugins/arcjet/skills/ from arcjet/skills main at f8959e141446b8f021a6b34e00211df9598a8e59 (skills#63). Published @arcjet/* 1.12.0 / Python arcjet 1.1.0: drop git pins, stamp Go v1.0.0-rc.2, mark remote policies MCP-only, and clarify adapter floors. Apply this repo's dprint formatter after the copy. Co-authored-by: David Mytton <davidmytton@users.noreply.github.com>
There was a problem hiding this comment.
Arcjet Review — 🟢 Low Risk
Decision: Checked
Rationale: Documentation-only sync of the vendored plugins/arcjet/skills/ tree to arcjet/skills main at f8959e14. All changes are to markdown files (SKILL.md, references/*.md) plus a CHANGELOG entry. The edits update version stamps (@arcjet/* 1.11.0 → 1.12.0, Python arcjet 1.0.0 → 1.1.0, Go v1.0.0-rc.1 → v1.0.0-rc.2), drop obsolete git-SHA install pins now that adapters ship on npm/PyPI, and add MCP Guard policy tool descriptions. No code, no config, no secrets, no auth/infra/CI/dependency-manifest changes. Diff is internally consistent and description accurately catalogs the changes.
Summary of Changes
Syncs the vendored plugins/arcjet/skills/ tree with arcjet/skills main at f8959e14: bumps version stamps to @arcjet/* 1.12.0 / Python arcjet 1.1.0 / Go v1.0.0-rc.2, replaces pip install ... @ git+... and @arcjet/guard SHA pins with published-package installs for Google ADK, TanStack AI, Claude Managed Agents, Claude Agent SDK (py), and Strands Agents (py), clarifies that remote actor/inputs support depends on installed adapter types, and documents MCP-only Guard policy tools. Adds a matching CHANGELOG entry. No behavior/code changes.
Notes
The AI assessed this PR as approvable, but the trust level (1) does not allow auto-approval. A human reviewer must approve this PR.
Review: cc6f53fe | Model: anthropic/claude-opus-4-7 | Powered by Arcjet Review
The worked-example move to arcjet/examples was intentional, but the directory on main is claude-managed-agent (singular). The plural path and the leftover arcjet-js/examples link both 404. Co-authored-by: David Mytton <davidmytton@users.noreply.github.com>
Summary
Updates the vendored skill tree to match current
arcjet/skillsmainatf8959e141446b8f021a6b34e00211df9598a8e59(arcjet/skills#63, merged 2026-09-16).That SHA is skills
mainafter the published@arcjet/*1.12.0 / Pythonarcjet1.1.0 sync (docs through arcjet-docs#939).This is a follow-up to #22, which synced
d9aadab3. Did not wait for open arcjet/skills#59 (Python ADK + Cloudflare Think) — that PR is still open (mergeable_statedirty) and is not onmain.Files were fetched from GitHub (cloned
arcjet/skillsat that SHA) rather than rewritten by hand. After copy, this repo’sdprintformatter was applied so CI’s format check passes.What changed
Canonical copy is
plugins/arcjet/skills/(skills/is the inbound symlink). Only that tree was updated. No new skill dirs and no newguards_js_*files.Version stamps and dropped git pins (skills#63)
@arcjet/*1.12.0 / Pythonarcjet1.1.0 / Gov1.0.0-rc.2pip install "arcjet[<extra>]")1.0.0+; Claude Agent SDK / Claude Managed Agents / Strands need 1.1.0list-guard-policies/describe-guard-policy/validate-guard-policy/put-guard-policy); the CLI has no policy commandsExisting files updated in place:
arcjet/SKILL.mdplus request/guard/MCP referencesguards_js_*.mdadapter files (unchanged:guards_js_claude_agent_sdk.md,guards_js_langgraph.md)integrate-arcjet-guard-*skill dirs (langchain-py,crewai,openai-agents-py,claude-agent-sdk-py,claude-managed-agents-py,strands-agents-py)Skills
mainstill has no JSintegrate-arcjet-guard-*dirs. JS adapters stay inarcjet/. HTTP frameworks stay in the request references.No extra remote-policy teaching beyond what that SHA already has (
actor/inputs/policyInputwording is from upstream).Follow-up on this branch: the Claude Managed Agents worked-example path is the verified singular
https://github.com/arcjet/examples/tree/main/examples/claude-managed-agent(plural andarcjet-jsexample links 404).What was left alone
Deprecated alias skill directories are not generated from the canonical skill. They are standalone deprecation stubs (
add-request-protection,add-guard-protection,protect-route,add-ai-protection) and were left unchanged.No marketplace catalog or plugin-manifest work. Hosts discover skills by scanning
plugins/arcjet/skills/(./skills/in the Codex manifest).Formatter note
dprinttable alignment treats unescaped||inside backticks as a column break (same issue as prior sync PRs). The Node version-range cells keep the escapes (\|\|) so the upstream ranges survive formatting:Node >=22.21.0 <23 || >=24.5.0(requests + guards JS tables)Test plan
dprint checkbash scripts/validate.shmainatf8959e14(arcjet/+ the sixintegrate-arcjet-guard-*dirs; existingguards_js_*.mdfiles; no JS integrate dirs on skills main)\|\|escapes survived dprint