Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add vasilistsavalias/conductor_for_codex --skill ai-repository-hygienegit clone --depth 1 https://github.com/vasilistsavalias/conductor_for_codexWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/vasilistsavalias/conductor_for_codex/ai-repository-hygiene)<a href="https://agentmods.dev/skills/vasilistsavalias/conductor_for_codex/ai-repository-hygiene"><img src="https://agentmods.dev/badge/skills/vasilistsavalias/conductor_for_codex/ai-repository-hygiene/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/vasilistsavalias/conductor_for_codex/ai-repository-hygiene"><img src="https://agentmods.dev/badge/skills/vasilistsavalias/conductor_for_codex/ai-repository-hygiene.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00109 | $0.01233 |
| Opus 5 | $0.00055 | $0.00616 |
| Sonnet 5 | $0.00022 | $0.00247 |
| Haiku 4.5 | $0.00011 | $0.00123 |
Grade A, and why
ai-repository-hygiene scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 109 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AI Repository Hygiene
Keep shared repositories professional while preserving private development material in an external sidecar. Treat this as company compliance, not as a history-concealment mechanism.
Non-negotiable boundaries
- Never rewrite Git history, force-push, delete branches, or remove worktrees as part of hygiene.
- Never move or delete a tracked path automatically, even when its name contains
agent,prompt,Gemini,Claude, orCodex. - Never treat product-owned runtime prompts, model code, RAG logic, evaluations, or AI documentation as private workspace debris.
- Never move
.env, credentials, keys, dependencies, databases, indexes, build outputs required in place, or configuration with a fixed runtime path. - Never discard dirty worktree changes.
- Never commit or push unless the user explicitly requests it.
- Never print secret contents. Report paths only.
- Compare paths case-insensitively against the complete Git index before moving. Windows can expose
gemini.mdand trackedGEMINI.mdas the same file.
Read references/policy.md when classification is ambiguous or when producing a compliance report.
Modes
Audit
Use audit by default. Make no changes.
python "$env:USERPROFILE\.codex\skills\ai-repository-hygiene\scripts\sidecar_hygiene.py" --repo .
The audit identifies safe sidecar candidates, tracked blockers, public ignore-file findings, protected paths, the inferred sidecar, and current Git status.
Apply
Use apply only when the user authorizes migration.
python "$env:USERPROFILE\.codex\skills\ai-repository-hygiene\scripts\sidecar_hygiene.py" --repo . --apply
Pass --sidecar-root <path> when the repository is not under a conventional github_repos/ directory. Apply mode moves only preflighted untracked root candidates and creates <repo>.code-workspace plus migration-report.json in the sidecar. It does not edit ignore files.
Workflow
- Record the repository root, branch,
git status --short, worktrees, and remotes. - Run the audit script. Do not infer safety from
.gitignoreor.git/info/exclude; ignored files can still be tracked or runtime-critical. - Inspect candidates using the policy classification:
- move private human/agent workspace artifacts;
- retain product-owned and runtime-required files;
- leave tracked policy violations in place and report them for an ordinary reviewed repository change;
- route secrets to approved secret storage, not the sidecar.
- Apply migration only after the candidate list is safe.
- Inspect remaining ignored/untracked files, including nested artifacts the root-candidate script intentionally does not move automatically.
- Clean ignore files narrowly:
- keep
.gitignorelimited to shared project outputs, environments, runtime data, and standard OS/editor files; - keep
.dockerignorelimited to files that must not enter the build context; - minimize
.git/info/excludeafter sidecar migration, retaining only genuine machine-local runtime paths.
- keep
- Before removing a Docker exclusion, inspect Dockerfiles and Compose build contexts. Git exclusions do not affect Docker.
- Verify behavior and repository integrity.
- Report current-tree compliance separately from historical findings. Do not imply that ordinary deletion removes old Git objects.
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 109 lines · 109 tokens per session scan A 994b2b67cfa0
ai-repository-hygiene is a skill published in the GitHub repository vasilistsavalias/conductor_for_codex (20 stars, last pushed 23d ago), licensed Apache-2.0. It adds 109 tokens to every session and 1,233 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
gh-issue
Fetch a GitHub issue, create a branch, implement with TDD, and open a PR.
ship-pr
Ship one pull request, monitor hosted CI and review feedback, follow PR policy, and request authorization before merging and cleaning up.
auto-skill-loop
Run an autonomous loop intended to result in merged PRs.
cut-release
Cut a new agnostic-ai release end to end. Use when the user wants to ship a new version (tag + GitHub Release).
skill-sync
Manage local skills backed by known remote Git repositories. Use when the user wants to import, inspect, diff, pull, resolve, push, reset, or remove them, or project current local skills to enabled clients with al sync. Do not use to discover new skills or create user-owned skills.
eg-precommit-review
Run a Codex-native pre-commit independent-review loop on pending changes.