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 rjwalters/kicad-tools --skill repo-helpgit clone --depth 1 https://github.com/rjwalters/kicad-toolsWrote 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/rjwalters/kicad-tools/repo-help)<a href="https://agentmods.dev/skills/rjwalters/kicad-tools/repo-help"><img src="https://agentmods.dev/badge/skills/rjwalters/kicad-tools/repo-help/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/rjwalters/kicad-tools/repo-help"><img src="https://agentmods.dev/badge/skills/rjwalters/kicad-tools/repo-help.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00023 | $0.01433 |
| Opus 5.5 | $0.00009 | $0.00573 |
| Sonnet 5.5 | $0.00005 | $0.00287 |
| Haiku 4.5 | $0.00002 | $0.00143 |
Grade B, and why
repo-help scanned grade B with 1 finding 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 2d 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.
Asks for rootmediumPrivilege escalation
A mod that escalates privileges can change anything on the machine, not only the project.
| $repo-sudo | One-time machine setup — grant passwordless sudo (validated drop-in) so an agent over SSH isn't blocked on password prompts | How it starts
The opening of the file, as written. The whole thing — 127 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Codex form of the
/repo:helpcommand. Other commands appear as$repo-<verb>skills. Helper scripts and.claude/paths below are relative to the repository root and exist when the Claude Code surface is installed alongside this one.
$repo-help — How to Use Repo Skills
Explain the repo skills installed in this repository. Introspect the actual install rather than assuming — installs can be filtered to a subset of commands.
Usage
$repo-help # Overview of all installed commands
$repo-help reset # Detailed help for one command
Steps — no argument (overview)
1. Read the actual install
ls .claude/commands/repo/*.md
cat .claude/skills/repo/install-metadata.json
cat .claude/skills/repo/.install-local.json 2>/dev/null # machine-local, gitignored
Build the command list from the files present. For each, pull the one-line
description: from its frontmatter. Do not list commands that are not
installed here.
install-metadata.json (tracked) carries version, commands, and dev. The
install date lives in the gitignored .install-local.json sidecar
(installed_at), so it is present only on the machine that ran the install;
older installs still keep installed_at inline in install-metadata.json. Use
whichever is present for the date below, and simply omit "(installed )"
when neither has it (e.g. a fresh clone on another machine) — its absence is
expected, not an error.
2. Present the overview
Structure it like this:
## Repo Skills v<version> (installed <date>)
General repository hygiene and environment tools. Commands **apply their
safe, reversible fixes by default** and report each change; add `--ask`
to review and confirm first. Irreversible removals (branches, worktrees,
stashes, untracked files) always need an explicit opt-in.
### Everyday
| Command | When to reach for it |
|---------|----------------------|
| $repo-reset | Done with a task — get back on main, synced, stale state reviewed |
| $repo-followups | End of session — file follow-on work surfaced this session as issues, here or in the right upstream tool repo |
| $repo-decide | Blocked on the operator — put the decision as ranked options (best to worst, each with why) they can answer with a number |
| $repo-handoff | Rolling the session (context full, CLI update) — preserve what only the session knows, then restart |
| $repo-tidy | Working tree cluttered with build artifacts and temp files |
| $repo-remote | Need a cloud dev box (GCP/AWS) with this repo ready to go |
| $repo-browser | Set up or check browser automation (Browser Use CLI, agent skill, auth, cloud credits) — report-first, installs only on confirm |
| $repo-sudo | One-time machine setup — grant passwordless sudo (validated drop-in) so an agent over SSH isn't blocked on password prompts |
| $repo-release | Cut a release — pre-flight, semver, CHANGELOG, version bump, tag, GitHub Release. Supports per-project release policy via named phase-boundary seams in `.repo/release-policy.md` |
### Periodic maintenance
| $repo-all | The whole hygiene pass in order — audit, scrub, docs, tidy, update-tools, deps check, reset |
| $repo-audit | Monthly sweep, or after a big refactor/import |
| $repo-scrub | What does the public surface expose? Code, history, issues, PRs, forks — report-only |
| $repo-host-optimize | Prep/re-check a Mac (or Linux box) for heavy Loom/agent build use — Gatekeeper churn, backup-agent interference, build-tree bloat |
| $repo-update-tools | Keep Loom/Anvil/Repo Skills installs current |
| $repo-deps | Keep third-party deps current — reconcile organization policy, Dependabot-first setup (Renovate opt-in), and bot PRs |
| $repo-optimize-ci | CI slow or expensive — find unfiltered jobs, broken cache keys, superseded runs; ranked by measured minutes saved |
| $repo-org-policy | Preview/install canonical organization preferences from the current client repo |
### Focused checks
| $repo-docs | Documentation health — content, README structure, cross-refs |
| $repo-branches | ... |
| $repo-gitignore | ... |
| $repo-links | ... |
| $repo-orphans | ... |
| $repo-readme | ... |
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.
- 2d ago Changed · +1 lines b24fda84808e
- 8d ago First seen · 126 lines · 23 tokens per session scan B f347c9dad214
repo-help is a skill published in the GitHub repository rjwalters/kicad-tools (73 stars, last pushed yesterday), licensed MIT. It adds 23 tokens to every session and 1,433 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it B with 1 finding (asks for root). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-11-01.
repo-help compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| repo-helprjwalters/kicad-tools | Skill | 73 | yesterday | MIT | B | 23 | Claude Code, Codex |
| ArXivdanielmiessler/LifeOS | Skill | 20k | 4d ago | MIT | A | 123 | Claude Code, Codex |
| live-tutorialleon-ai/leon | Skill | 18k | yesterday | MIT | A | 36 | Claude Code, Codex |
| Explaindanielmiessler/LifeOS | Skill | 20k | 4d ago | MIT | A | 0 | Claude Code, Codex |
| founder-coachalirezarezvani/claude-skills | Skill | 30k | 4d ago | MIT | A | 41 | Claude Code |
| cs-onboardalirezarezvani/claude-skills | Skill | 30k | 4d ago | MIT | A | 47 | Claude Code |
| study-plangenspark-ai/genoffice | Skill | 12k | today | Apache-2.0 | A | 30 | Claude Code, Codex |