## Summary:
- add experimental gitlab-address-comments skill docs and workflow steps
- include a fetch_comments.py helper to pull MR discussions via glab api
- add Apache 2.0 license for the new skill bundle
## Testing:
- manually QA'd
## Overview
This PR adds two Codex readiness skills that evaluate repository
guidance quality
and end‑to‑end agentic execution. The unit test focuses on deterministic
checks + in‑session LLM evaluation of AGENTS.md/PLANS.md quality, while
the
integration test runs a full agentic loop and scores real code changes
and
build/test outcomes.
## 1) codex-readiness-unit-test (LLM Codex Readiness Unit Test)
### Goal
Validate that AGENTS.md and PLANS.md provide sufficient, usable guidance
using
deterministic checks plus in‑session LLM evaluation, and generate a
scored
JSON + HTML report.
### How it works
This skill builds a report from two pipelines: deterministic filesystem
checks
and in‑session LLM evaluation of AGENTS.md/PLANS.md guidance. It writes
a
timestamped run directory with evidence, LLM results, and a scored
JSON/HTML
report; in optional execute mode it runs a user‑approved plan and
includes
execution logs in scoring. JSON outputs are strictly validated with a
retry +
json‑fix loop.
## 2) codex-readiness-integration-test (LLM Codex Readiness Integration
Test)
### Goal
Validate real agentic execution quality by running Codex CLI against the
repo,
executing an approved change prompt, and scoring results with evidence +
LLM
evaluation.
### How it works
This skill runs an end‑to‑end agentic execution against the repo using
Codex
CLI, then executes a build/test plan and scores the run from evidence
plus LLM
evaluation. It spins up the Codex session by launching the CLI as a
subprocess
with HOME/XDG_CACHE_HOME pointed at the repo‑local .codex-home, using
the
approved prompt.json (change prompt + agentic_loop settings) so the CLI
reads
AGENTS.md and operates in the repo. It requires a repo‑local login,
always
runs in execute mode, and writes results to a timestamped run directory
with
agentic logs, LLM results, and a summary.
## Summary
This PR corrects the path to the CLA document referenced in the GitHub
Actions workflow. The CLA document is located at the root of the
repository, not in a docs/ subdirectory.
The CLA and CLA workflow were copied verbatim from the openai/codex
repo. The contributing guidelines are a stripped-down version of that in
the openai/codex repo.
---------
Co-authored-by: Josh McKinney <joshka@openai.com>
Rewrites `plan` skill as `create-plan` and removes all lifecycle
management skills. Moved to a new `.experimental` bucket for evaluation
and to get feedback on use.
This PR removes references to oai_gh from the curated GitHub skills and
replaces them with standard GitHub CLI auth guidance (gh auth login).
The goal is to avoid assuming internal tooling in a public repo.
That said, I'm not sure whether oai_gh is actually a built-in Codex
helper. If it is, feel free to close this PR.
I also did a sweep of .curated for other internal-only helpers and
didn't find any additional ones.
- Install curated skills from openai/skills or other repos.
- Named it "skill-installer" to fit with "skill-creator".
- We'd discussed installing to `$CODEX_HOME/skills/.curated`, but now
that the skill can also install non-curated skills, I think we should
just install to `$CODEX_HOME/skills`.
- Need to test on Windows.
Vaibhav SrivastavandGitHub
LukeandGitHub
Eric TrautandGitHub
ErayandGitHub
Josh McKinneyandGitHub
xl-openaiandGitHub
Gav VermaandGitHub
Christian Bager Bach HoumannandGitHub
aeandGitHub
