everything-22d18f — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited everything-22d18f (Plugin) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Write a skill once, run it in every coding agent. Portable SKILL.md skills for Claude Code, GitHub Copilot CLI, Codex, and Cursor — from one source of truth.
Portable Agent Skills by Biswajeet Das. One SKILL.md source of truth, installable across Claude Code, GitHub Copilot CLI, Codex CLI, and Cursor via per-harness manifests — because skills are an open standard, not a single-vendor format.
| Skill | What it does |
|---|---|
| [reviewloop](skills/reviewloop/SKILL.md) | Drives a PR to all-clear across every reviewer — Greptile, CodeRabbit, Copilot, other bots, and human teammates. Bots get re-triggered and re-checked in a loop; humans get one pass (addressed + re-requested) then it hands back. The reviewer-agnostic answer to "clear all the reviews on this PR." |
| [ciloop](skills/ciloop/SKILL.md) | Drives a failing CI run back to green. Pulls the real error from the GitHub Actions logs, reproduces it locally (so it's not burning CI minutes per attempt), fixes the failures it's confident about, pushes once, and re-watches — looping until green. Hands back flaky/infra failures instead of flailing. The red-CI twin of reviewloop. |
| [standup](skills/standup/SKILL.md) | Generates a paste-ready daily standup from your GitHub activity — merged, opened, and reviewed PRs across the whole org — rendered as a Done / Today / Blockers block. Weekday-aware window (Monday reaches back to Friday). No more "what did I do yesterday?" |
Install once — the skills plugin bundles all three.
This repo is a marketplace with one plugin (`skills`) that bundles all three skills. A Claude Code plugin loads every skill it contains, so a single install gets the whole set — cleanly, no duplicates.
claude plugin marketplace add BiswaViraj/agent-skills
claude plugin install skills@biswaviraj-skillsOr, with @biswaviraj/cc-setup, it's in the palette — run npx @biswaviraj/cc-setup@latest and pick it.
Copilot CLI shares Claude Code's plugin/marketplace system:
copilot plugin marketplace add BiswaViraj/agent-skills
copilot plugin install skills@biswaviraj-skillscodex plugin marketplace add BiswaViraj/agent-skills
codex plugin install skills@biswaviraj-skillsOr drop a single skill straight into a folder Codex scans (~/.agents/skills/ for personal, <repo>/.agents/skills/ per-project):
git clone https://github.com/BiswaViraj/agent-skills /tmp/agent-skills
cp -r /tmp/agent-skills/skills/standup ~/.agents/skills/standupcursor plugin marketplace add BiswaViraj/agent-skills
cursor plugin install skills@biswaviraj-skillsreviewloop — on a branch with an open PR:
reviewloop this PR
or "clear all the reviews", "address every reviewer", "loop until the PR is clean".
ciloop — when CI is red:
ciloop
or "fix the failing CI", "make the checks pass", "loop until CI is green".
standup — anytime:
write my standup
or "what did I do since Friday", "what did I ship this week".
agent-skills/
├─ .claude-plugin/ # Claude Code + Copilot CLI — marketplace.json (one 'skills' plugin)
├─ .codex-plugin/ # Codex CLI (plugin.json)
├─ .cursor-plugin/ # Cursor (plugin.json)
└─ skills/
├─ reviewloop/
│ ├─ SKILL.md
│ └─ references/github-mechanics.md
├─ ciloop/
│ ├─ SKILL.md
│ └─ references/gh-ci.md
└─ standup/
├─ SKILL.md
└─ standup.shAdding a skill: drop a new folder under skills/ — it ships with the skills plugin automatically (a plugin loads every skill it contains; no marketplace.json change needed).
The harnesses these skills run in: Claude Code · GitHub Copilot CLI · Codex CLI · Cursor
The review bots reviewloop drives: Greptile · CodeRabbit · GitHub Copilot code review
The standard: Agent Skills — the open SKILL.md spec adopted across 30+ agents.
See also: @biswaviraj/cc-setup — one command to load your plugins + MCP servers (including these skills) into any project.
<sub>Keywords: claude code skills · agent skills · SKILL.md · github copilot cli plugins · codex cli skills · cursor skills · automated pr review · greptile / coderabbit / copilot review loop · ai code review automation · auto-fix failing CI · github actions fix loop · automated standup generator · anthropic agent skills marketplace</sub>
MIT © Biswajeet Das
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.