feat(adapters): codex T3 — skills mirroring the vouch slash commands#394
Merged
Conversation
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Plus Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
github-actions
Bot
added
adapters
This was referenced Jul 6, 2026
Open
7 tasks
There was a problem hiding this comment.
Pull request overview
Adds Codex Tier-3 support so Codex users get the same guided vouch workflows that Claude Code already has, using project-local Codex skills to preserve the “no home-directory writes” invariant. The PR also expands the adapter installer to support Codex-specific integration needs (fenced snippet refresh + TOML deep-merge) and adds tests that lock these surfaces in sync.
Changes:
- Add Codex T3 project-local skills under
.codex/skills/, reusing the existing OpenClaw skill bodies and sync-testing them against Claude Code commands. - Extend
install_adapterwithtoml_mergesupport (for.codex/config.toml) and refresh-in-place behavior for fenced snippets. - Add Codex adapter docs + AGENTS snippet and broaden installer tests to cover T2/T3/TOML merge invariants.
Reviewed changes
Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.
Show a summary per file
| File | Description |
|---|---|
src/vouch/install_adapter.py |
Adds toml_merge install strategy and refresh-in-place fenced snippet behavior. |
tests/test_install_adapter.py |
Adds coverage for Codex T2/T3 installs, scope checks, idempotency, and body sync. |
adapters/codex/install.yaml |
Declares Codex tiers: T1 toml-merge config, T2 fenced AGENTS snippet, T3 skills installation. |
adapters/codex/README.md |
Documents Codex T3 skills behavior and rationale vs global prompts. |
adapters/codex/AGENTS.md.snippet |
Adds Codex-specific AGENTS snippet (kept in lockstep with Cursor). |
Comment on lines
+145
to
+152
| fenced = bool(raw.get("fenced_append", False)) | ||
| json_merge = bool(raw.get("json_merge", False)) | ||
| toml_merge = bool(raw.get("toml_merge", False)) | ||
| parsed_entries.append( | ||
| _FileEntry(src=src, dst=dst, fenced_append=fenced, json_merge=json_merge) | ||
| _FileEntry( | ||
| src=src, dst=dst, fenced_append=fenced, | ||
| json_merge=json_merge, toml_merge=toml_merge, | ||
| ) |
Comment on lines
+23
to
+31
| ## Skills (T3) | ||
|
|
||
| `vouch install-mcp codex --tier T3` also drops the vouch guided flows | ||
| (`vouch-recall`, `vouch-status`, `vouch-resolve-issue`, | ||
| `vouch-propose-from-pr`, plus the company-brain set) into the | ||
| project-local `.codex/skills/` directory. Codex discovers skills from | ||
| `<project>/.codex/skills/` in trusted projects, so they surface in the | ||
| session automatically — ask for a skill by name (e.g. "use the | ||
| vouch-recall skill for X") or let codex pick them up from context. |
Comment on lines
+471
to
+476
| for name in _CODEX_SKILL_NAMES: | ||
| text = (tmp_path / ".codex" / "skills" / name / "SKILL.md").read_text( | ||
| encoding="utf-8" | ||
| ) | ||
| assert f"name: {name}" in text.split("---", 2)[1], name | ||
|
|
vouchdev#361 inserted `_log = logging.getLogger(...)` between two import groups in jsonl_server.py. that statement-among-imports trips ruff's E402 on every import that follows it, so `ruff check src tests` — the lint gate in ci — now fails on the whole repo. move the logger definition below the imports (where storage.py already keeps its own module logger); no behaviour change.
.codex/config.toml is codex's primary config file, so the plain-copy path silently skipped any project where codex was already configured and vouch never got wired. add a toml_merge entry flag mirroring json_merge: parse the existing destination with tomllib, deep-merge the template's tables into it (existing user values always win on conflict), and write back. flip the codex T1 entry to toml_merge. writing uses a minimal hand-rolled serializer (tables, arrays, inline tables in arrays, scalars, datetimes) so the dependency set stays unchanged; the output must survive a tomllib round-trip back to the merged data, and anything the serializer can't faithfully re-emit degrades to skipped rather than risking the user's config. closes vouchdev#384
codex reads AGENTS.md for project instructions the way cursor does, but the codex adapter stopped at T1, so a codex session got the kb tools with no standing guidance on recall-first or the review gate. ship adapters/codex/AGENTS.md.snippet as a T2 tier with the standard fence markers, kept in lockstep with cursor's snippet modulo the host name (enforced by a sync test). also close the edited-fence gap in _install_fenced per the ticket's acceptance criteria: a fence body that drifted from the shipped snippet is replaced within the markers (reported as merged) instead of being skipped forever, while user content outside the fence stays untouched. a begin marker without an end marker is treated as corrupt and left alone. closes vouchdev#385
claude-code T3 ships nine slash commands; codex users got none of the guided flows. ship them as codex skills under the project-local .codex/skills/ as a T3 tier. scope decision the ticket asked to resolve first: codex loads custom prompts only from ~/.codex/prompts/ (user-global) and has deprecated them upstream in favour of skills, which do have a project-local home at <project>/.codex/skills/ in trusted projects. the vouchdev#179 rule forbids a project-scoped install from touching home-directory state, so prompts are out and skills are in — recorded in the manifest comment and the adapter readme, with a manual-copy pointer for users who still want ~/.codex/prompts. the SKILL.md files are referenced from the openclaw mirror rather than duplicated (same reuse pattern openclaw itself applies to the claude-code commands), so one edit updates every host. a parametrized sync test asserts each installed codex skill body equals the matching claude-code command body, and a scope test asserts every installed path stays under the project target. closes vouchdev#386
dripsmvcp
force-pushed
the
feat/codex-skills
branch
from
91bd494 to
109a3a3
Compare
Contributor
Author
|
addressed the copilot review on both points: the skill-name test now matches the whole |
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 free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
claude-code T3 ships nine slash commands (the four core flows plus the company-brain set); codex users got none of them, so the guided flows were claude-code-only. this ships them to codex as a T3 tier.
the scope decision the ticket asked to resolve first: checked against the current codex cli — custom prompts load only from
~/.codex/prompts/(user-global, no project-local location was ever added) and are deprecated upstream in favour of skills. skills do have a project-local home at<project>/.codex/skills/in trusted projects. since the #179 rule forbids a project-scoped install from touching home-directory state, prompts are out and project-local skills are in. the choice is recorded in the manifest comment as the ticket requested, and the adapter readme documents how the surface appears inside codex plus a manual-copy pointer for anyone who still wants~/.codex/prompts/.the SKILL.md files are referenced in place from the openclaw mirror rather than duplicated — the same reuse pattern openclaw's own manifest applies to the claude-code commands — so there is exactly one copy of each skill body in the repo and one edit updates every host. openclaw's existing sync test already pins those bodies to the claude-code commands; a new parametrized test asserts each installed codex skill body equals the matching claude-code command body, closing the loop from the codex side.
tests also cover: all nine skills land at T3, frontmatter names match their directory names (codex resolves skills by name), idempotent re-run, and a #179 scope check that every installed path stays under the project target. verified end-to-end with
vouch install-mcp codex --tier T3into a temp project — 11 files written (config + snippet + 9 skills), all under the target.stacks on #392 and #393 (shared
adapters/codex/install.yaml) — merge after them; the diff reduces to the T3 change once they land.closes #386