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 PaulRBerg/agent-skills --skill refresh-cligit clone --depth 1 https://github.com/PaulRBerg/agent-skillsWrote 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/paulrberg/agent-skills/refresh-cli)<a href="https://agentmods.dev/skills/paulrberg/agent-skills/refresh-cli"><img src="https://agentmods.dev/badge/skills/paulrberg/agent-skills/refresh-cli/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/paulrberg/agent-skills/refresh-cli"><img src="https://agentmods.dev/badge/skills/paulrberg/agent-skills/refresh-cli.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 | $0.00029 | $0.01380 |
| Opus 5.5 | $0.00012 | $0.00552 |
| Sonnet 5.5 | $0.00006 | $0.00276 |
| Haiku 4.5 | $0.00003 | $0.00138 |
Grade A, and why
refresh-cli-skill 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 17d 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 — 95 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Refresh CLI Skill
Refresh skills/cli-* when an installed CLI is newer than its references/version.txt marker. Review
skills/chromium-browser against Chrome DevTools MCP every week, even when the installed version has not changed.
Circadian's weekly maintenance job runs jobs/refresh-cli-skills.sh in an isolated clone of this repository. The job
commits, pushes, and publishes the result after validation. It can also request guidance reviews for explicitly named
skills whose watched upstream sources changed.
Arguments
- Accept arguments as
<skill>=<version>, for examplecli-just=1.43.0orchromium-browser=1.10.1. - Treat each version as an already-normalized semver with no leading
v. - Accept version pairs only for
cli-*andchromium-browserskills. - Also accept explicit upstream review names and source notes supplied alongside the pairs. The scheduler says
Also review <space-separated names> against themand appends notes fromSKILL_REFRESH_UPSTREAM_NOTES. - Confirm that every additional name has an existing
skills/<name>/SKILL.md. Reject unknown names. - Upstream-only reviews need no version pair or
references/version.txtmarker. Do not invent either. - If neither valid skill/version pairs nor explicit upstream review names are supplied, stop.
Workflow
- Work in the current checkout. In an isolated clone, never touch the shared
~/projects/agent-skillsworktree. - Before editing, confirm the worktree is clean. If it is dirty, stop and report the dirty paths.
- For each requested CLI skill:
- Confirm
skills/<cli-skill>/SKILL.mdexists. - Map
cli-<name>to binary<name>. Mapcli-coingeckotocg. - Confirm the binary's
--versionoutput still reports the requested version. If evidence is missing, stop with failure. - Read current docs and references that mention versioned features, flags, or upstream links.
- Inspect official stable release notes and tagged manuals or CLI docs for the requested version.
- Use installed CLI help to confirm compatibility.
- Update stale or missing facts. Keep edits terse and preserve local safety rules.
- Write
skills/<cli-skill>/references/version.txtwith exactly the requested semver and one trailing newline.
- Confirm
- For
chromium-browser, complete these steps even when the installed version has not changed:- Confirm
skills/chromium-browser/SKILL.mdexists. - Read the installed wrapper
~/.agents/skills/chromium-browser/scripts/chrome-devtools-mcp.shwithout changing it. Its source isskills/chromium-browser/scripts/chrome-devtools-mcp.shin this checkout. - Verify the wrapper's installed executable version through package metadata or its
--versionoutput. Never start the MCP server for this check. - Inspect official stable Chrome DevTools MCP release notes and tagged documentation. Compare them with the installed version and wrapper flags.
- Update compatible skill guidance and link to the installed release's tagged tool reference. No browser version marker is required.
- Report stable upstream releases newer than the installed version and wrapper incompatibilities you could not resolve. Never present unreleased or unavailable features as available.
- Preserve attach-only operation, privacy and header redaction, negotiated-root restrictions, screenshot caps, and shared-browser ownership.
- When the installed executable's
--helpshows that a wrapper flag was removed, renamed, or became the default, or that an environment variable the wrapper sets changed, update the source wrapper andtests/chromium-browser/wrapper.batsto match the installed release, then runjust bats-test tests/chromium-browser. Otherwise leave the wrapper unchanged. Never change settings, client configuration, or installed copies.
- Confirm
- For each additional upstream review name:
- Read its source notes and affected skill guidance. Inspect every supplied source and the relevant stable changes.
- Check official documentation and source evidence before changing guidance. Confirm installed compatibility when the proposed guidance depends on an installed tool or dependency.
- Change only guidance affected by those sources. Preserve local behavior and safety rules.
- Do not update dependencies or merge external skill forks through this review. Follow
@rebase-skillonly under its separate invocation policy. - Do not create or advance a version marker for an upstream-only review.
- Edit only the requested skill directories and, for a wrapper update,
tests/chromium-browser/. Never edit generated lockfiles, unrelated skills, installed copies under~/.agents, or global Claude/Codex configuration. - Run
just prettier-write <changed files> --ignore-unknown. Then runjust skill-checkandjust prettier-check skills/<name> ...for every selected skill directory. If a selected directory has an existing formatting failure, format and validate the affected file within that directory. Do not report success while a selected-directory check fails. Complete the local gate for any changed helper or behavior. - If the scheduled weekly job invoked you, stop without committing or publishing. The job performs both steps.
Otherwise, commit and publish per
AGENTS.md. - Report evidence sources, compatible documentation changes, newer upstream releases, and configuration incompatibilities. If a required check or evidence source fails, report failure instead of claiming success.
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.
- 17d ago Changed · +3 lines e82ad328eff3
- 20d ago Changed · +2 lines ee88b8619640
- 24d ago Changed · +13 lines 3c9e0f14c224
- 27d ago Changed · +1 lines 21f552f5d185
- 1mo ago Changed · +26 lines 3a529c19415c
- 1mo ago First seen · 50 lines · 29 tokens per session scan A e58631fff158
refresh-cli-skill is a skill published in the GitHub repository PaulRBerg/agent-skills (104 stars, last pushed 2d ago), licensed MIT. It adds 29 tokens to every session and 1,380 once invoked, about $0.0001 per session on Opus 5.5. 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 2026-10-06.
refresh-cli-skill compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| refresh-cli-skillPaulRBerg/agent-skills | Skill | 104 | 2d ago | MIT | A | 29 | Claude Code, Codex |
| imagegenopenai/codex | Skill | 134k | today | Apache-2.0 | A | 113 | Codex |
| local-ai-agentsmicrosoft/ai-agents-for-beginners | Skill | 79k | 2d ago | MIT | A | 200 | Claude Code, Codex |
| playwright-climicrosoft/playwright | Skill | 100k | yesterday | Apache-2.0 | A | 19 | Claude Code |
| browser-automationopenclaw/openclaw | Skill | 394k | today | MIT | A | 36 | Claude Code |
| systematic-debuggingobra/superpowers | Skill | 315k | 5d ago | MIT | A | 21 | Claude Code |
| next-cache-components-adoptionvercel/next.js | Skill | 143k | today | MIT | A | 95 | Claude Code |