agentsclimarketplace

Repo status

Skill VictorAurelius/claude-starter-kit/skills/workflow/repo-status

Use when the user says 'status', 'repo status', 'tình trạng repo', 'health check', 'is the repo OK?', 'security', 'CVE', 'vuln', or when starting a conversation that needs a quick remote-repo assessment. Checks: CI, PRs/branches, audit gaps, GitHub Security (Dependabot + code-scanning + secret-scanning) → output level GREEN/YELLOW/ORANGE/RED/BLACK.From its SKILL.md

Install
npx -y skills add VictorAurelius/claude-starter-kit --skill repo-status

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
  • runs commandsInstructs the agent to run 8 commands, including `./scripts/repo-status.sh` and 7 more.

SKILL.md

6.3 KB, ~1.6k tokens by cl100k_base, as published. Nobody here has run it

/repo-status — Remote Repo Health Check

Usage: /repo-status or /repo-status --quick

Assess remote-repo health across 4 factors, output a status level.


Instructions

Step 1: Gather data

If the project provides a helper script scripts/repo-status.sh, use it:

./scripts/repo-status.sh          # full report
./scripts/repo-status.sh --json   # JSON for parsing
./scripts/repo-status.sh --level  # level only (quick check)

If no such script exists, gather the 4 factors directly with gh / git (Steps below). If --quick: report the level in one line and stop.

Step 2: Analyze the 4 factors

Factor 1 — CI: if CI is failing on the default branch, find root cause:

DEFAULT=$(git symbolic-ref --quiet --short refs/remotes/origin/HEAD | sed 's@^origin/@@'); DEFAULT=${DEFAULT:-main}
FAILED_RUN=$(gh run list --branch "$DEFAULT" --limit 5 \
  --json databaseId,workflowName,conclusion \
  --jq '.[] | select(.conclusion=="failure") | .databaseId' | head -1)
gh run view "$FAILED_RUN" --log-failed 2>/dev/null | tail -30

Factor 2 — PRs/Branches: check whether stale branches are WIP or abandoned:

git for-each-ref --sort=-committerdate \
  --format='%(refname:short) %(committerdate:relative)' refs/remotes/ \
  | grep -vE "main|master|HEAD"

Factor 3 — Audit Gaps (optional — only if the project runs audits): read the latest audit report under documents/04-quality/audits/** (if present) to understand gap context. Cross-check: do gaps already have a merged PR fix? (compare against gh pr list --state merged since the audit date). If the project has no audit pipeline, skip this factor.

Factor 4 — GitHub Security: 3 endpoints via gh api:

REPO=$(gh repo view --json nameWithOwner --jq .nameWithOwner)
gh api "repos/$REPO/dependabot/alerts?state=open"       # → 403 if disabled
gh api "repos/$REPO/code-scanning/alerts?state=open"    # → [] if no scanner
gh api "repos/$REPO/secret-scanning/alerts?state=open"  # → public/enterprise only

Severity mapping:

  • Dependabot critical → BLACK; high → RED; medium → YELLOW
  • Code scanning error → HIGH (counts as CVE); warning → MEDIUM; note → LOW
  • Any secret-scanning alert → BLACK (credential exposure)
  • Dependabot disabled (HTTP 403) → minimum ORANGE (silent-drift risk)

If Dependabot is disabled, prompt the user to enable it (repo Settings → Code security and analysis).

Step 3: Output Report

## Repo Status: [LEVEL] — [Label]

**Date:** [date]
**Branch:** [default] @ [commit hash]

### Factor 1: CI
- Status: ✅/❌ [details]
- Days since green: [N]
- Root cause (if fail): [description]

### Factor 2: PRs & Branches
- Open PRs: [N]
- Stale branches: [N]
- Details: [list]

### Factor 3: Audit Gaps   (omit if no audit pipeline)
- Latest audit: [date] — [score]
- Unfixed P0/P1/P2: [counts]

### Factor 4: GitHub Security
- Dependabot: enabled/disabled — [critical/high/medium/low counts]
- Code scanning: [N errors / N warnings / N notes] — top finding: [rule.id + file]
- Secret scanning: [N open alerts] (BLACK if >0)

### Recommended Actions
1. [Highest priority]
2. [Next]

Step 4: Recommend actions

LevelAction
GREENContinue normal development
YELLOWNote minor items, fix when convenient
ORANGEFix before opening new PRs
REDFix now — top priority
BLACKStop all work, fix CI/security first

Level Definitions

Detail: reference/level-definitions.md

LevelLabelTrigger
GREENHealthyCI green + 0 open PRs/stale branches + 0 P0/P1 gaps + 0 CVE + Dependabot enabled
YELLOWMinor IssuesCI green + minor gaps (P2/P3) OR 1-2 stale branches OR code-scan warnings
ORANGENeeds AttentionCI green + P1 gaps, OR >2 stale items, OR Dependabot disabled, OR ≥3 warnings
REDDegradedCI red on default branch OR unfixed P0 gaps OR ≥1 HIGH CVE
BLACKBrokenCI red >7 days OR CRITICAL CVE OR secret-scanning alert

Rules

  • Run the helper script first if it exists — don't infer status by hand.
  • LUÔN giao tiếp tiếng Việt (per kit bilingual convention).
  • If a check fails (GitHub API unavailable) → report it, don't guess.
  • Cross-reference audit gaps with merged PRs — a gap may already be fixed but not re-audited.
  • Recommended actions must be actionable (PR scope, estimate).

Gotchas

  • All checks need gh authenticated (gh auth status).
  • ci_days_red is based on last-success date, not first-failure date.
  • Audit gaps (Factor 3) parse via P0/P1/P2 markers — if a project's report format differs, adapt the parse.
  • Stale branches = unmerged into default branch, may be valid WIP — check commit date before recommending delete.
  • Security factor (F4): gh api security endpoints return 403 if Dependabot disabled. Use a jq type-check (type=="array" = enabled) to distinguish a disabled object from an empty-but-enabled array. Do NOT grep the response body — CVE descriptions may contain the word "disabled" and cause false positives.
  • Secret-scanning endpoint is only available for public repos or GitHub Enterprise — returns 404 for private repos → treat as "disabled".
  • Code-scanning alerts only populate when a CodeQL/Trivy/SARIF upload workflow runs. No scanner = empty [], not disabled.

Skill contents

  • SKILL.md — this file
  • reference/level-definitions.md — 5-level definitions + decision matrix + per-factor detail

Related

  • .claude/skills/workflow/start-session/SKILL.md — faster routine entry (consumes repo level)
  • .claude/rules/mcp-first-with-fallback.md — prefer GitHub MCP over gh CLI when connected
  • .claude/rules/output-review-mandate.md — health reports are reviewable outputs

What ships with it: 1 file

5.7 KB alongside SKILL.md

reference/

Gives 0 of the 12 instructions most vulnerability scanning skills give in ~1.6k tokens

Counted across 223 of the 238 authors here whose files we hold, read 2026-09-06

  • Fix critical findings immediatelyin 21 of 223, across 14 files
  • Fix high findings before productionin 18 of 223, across 12 files
  • Check and install AgentShield before scanningin 18 of 223, across 12 files
  • Scaffold a secure configuration with initin 16 of 223, across 10 files
  • Add the AgentShield GitHub Action to CIin 16 of 223, across 10 files
  • Run the three-agent opus pipeline for deeper analysisin 14 of 223, across 8 files
  • Apply safe auto-fixes onlyin 14 of 223, across 8 files
  • Use JSON output for CI/CD integrationin 13 of 223, across 7 files
  • Filter findings with a minimum severityin 12 of 223, across 10 files
  • Classify each finding by severityin 10 of 223
  • Use parameterized queries for all database accessin 10 of 223, across 9 files
  • Write tests before writing the rulein 9 of 223, across 4 files

Said here and by no other author read

  • Run the helper script first if it exists
  • Report the level in one line for --quick
  • Find the root cause when default-branch CI fails
  • Check whether stale branches are WIP or abandoned
  • Skip the audit factor when no audit pipeline exists
  • Cross-check audit gaps against merged PRs

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.