wip-ai-devops-toolbox@1.9.72
CAUTION· startup onlyPublisher claims: "Complete DevOps toolkit for AI-assisted software development. Release pipeline, license compliance, copyright enforcement, repo visibility guard, identity fi..." Observed in sandbox: no network, touches nothing outside the project, writes no memory.
Coverage
Startup only; authenticated tool behavior not tested.- Tested
- server startup over stdio (initialize sent) · every file read, file written and host contacted by what ran
- Not tested
- the tool list (the server exited before answering) · what the tools do when called with real credentials
Is [email protected] safe to install?
KillerSkills ran it in the KillerSkills honeypot sandbox (2 scenarios; startup only; authenticated tool behavior not tested) on 2026-11-20 and scored it 64/100 (Caution). It is not known to be malicious, but it has behaviour or code a security policy should decide on before install. Notable findings: HTML comment carrying imperative instructions to the agent; Destructive command against a filesystem root or device; Destructive command against a filesystem root or device. Summary: no network, touches nothing outside the project, writes no memory.
Behavior trace
Honeypot run · 21 s · 2 scenarios · 0 deviationsStatic findings
Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.# docker rm -f / docker system pruneFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.# rm -rf / rm -r / rm --recursiveFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.3. If `bun` is not installed: `curl -fsSL https://bun.sh/install | bash`Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.1. `/codex review` or `/codex review <instructions>` — **Review mode** (Step 2A)Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.wip-readme-format /path/to/repo --check # validate existing README against standardFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.codex exec resume <session-id> "<prompt>" -s read-only -c 'model_reasoning_effort="high"' --enable web_search_cached --json 2>"$TMPERR" | python3 -c "Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.wip-readme-format /path/to/repo --deploy # assemble sections into README.mdFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.wip-readme-format /path/to/repo --dry-run # preview without writingFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.| `rm -rf` / `rm -r` / `rm --recursive` | `rm -rf /var/data` | Recursive delete |Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.wip-readme-format /path/to/repo # generate README-init-*.md section filesFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.codex exec "<prompt>" -s read-only -c 'model_reasoning_effort="high"' --enable web_search_cached --json 2>"$TMPERR" | python3 -c "Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.3. If \`bun\` is not installed: \`curl -fsSL https://bun.sh/install | bash\``;Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.codex exec "<prompt>" -s read-only -c 'model_reasoning_effort="xhigh"' --enable web_search_cached --json 2>/dev/null | python3 -c "Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.Regenerate: bun run gen:skill-docsFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.3. If `bun` is not installed: `curl -fsSL https://bun.sh/install | bash`Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.- **Auto-update** — Factual corrections clearly warranted by the diff: adding an item to aFix: Pin versions and sources; vendor anything fetched at install time.…ee. This is critical — without it, screenshots are invisible to the user. 12. **Never refuse to use the browser.** When the user invokes /qa or /qa-only, they are requestin…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.…ee. This is critical — without it, screenshots are invisible to the user. 12. **Never refuse to use the browser.** When the user invokes /qa or /qa-only, they are requestin…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.GATE: PASS Tokens: 14,331 | Est. cost: ~$0.12Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.3. If `bun` is not installed: `curl -fsSL https://bun.sh/install | bash`Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.3. If `bun` is not installed: `curl -fsSL https://bun.sh/install | bash`Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.rm -f /tmp/gstack-pr-body-$$.mdFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.### B7. Update CLAUDE.mdFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.- Always accept the user's final choice. Never refuse to proceed.Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.- A) Ship it — write DESIGN.md and CLAUDE.mdFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.3. If `bun` is not installed: `curl -fsSL https://bun.sh/install | bash`Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.### B7. Update CLAUDE.mdFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.3. If `bun` is not installed: `curl -fsSL https://bun.sh/install | bash`Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.# docker rm -f / docker system pruneFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.# rm -rf / rm -r / rm --recursiveFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.**Update CLAUDE.md** (or create it if it doesn't exist) — append this section:Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.expect(() => validateNavigationUrl('http://0xA9FEA9FE/')).toThrow(/cloud metadata/i);Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.expect(() => validateOutputPath('/tmp/../etc/passwd')).toThrow(/Path must be within/);Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.…e/MEMORY.md`, `workspace/IDENTITY.md`, `workspace/SOUL.md`, `workspace/WHERE-TO-WRITE.md`, `workspace/HEARTBEAT.md`, `workspace/memory/*.md`, `CLAUDE.md`. Worktree convention: `.worktrees/<repo>--<branch>/`. Bootstrap compound (`gi…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.3. If `bun` is not installed: `curl -fsSL https://bun.sh/install | bash`Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.Run full audit, then load previous `design-baseline.json`. Compare: per-category grade deltas, new findings, resolved findings. Output regression table in report.Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.3. If `bun` is not installed: `curl -fsSL https://bun.sh/install | bash`Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.test_case "rm on main-tree path still denies" deny Bash "rm /some/repo/file.md" trueFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.// Force kill browser instead of graceful close (avoids hang)Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.console.log(`FRESH: ${relOutput}`);Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.11. **Wartime awareness** — Correctly diagnose peacetime vs wartime. Peacetime habits kill wartime companies (Horowitz).Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.wip-readme-format /path/to/repo # rewrite README.mdFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.console.log(`GENERATED: ${relOutput}`);Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.await handleWriteCommand('cookie-import', ['../../etc/shadow'], bm);Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.await handleReadCommand('eval', ['../../etc/passwd'], bm);Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.console.log(` [check] \u274c ${output} (${totalValid} valid)`);Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.3. If \`bun\` is not installed: \`curl -fsSL https://bun.sh/install | bash\``;Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.await handleWriteCommand('cookie-import', ['/etc/passwd'], bm);Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.3. If `bun` is not installed: `curl -fsSL https://bun.sh/install | bash`Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.// wip-readme-format /path/to/repo # generate README-init-*.md section filesFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.// wip-readme-format /path/to/repo --deploy # assemble sections into README.md + TECHNICAL.mdFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.// wip-readme-format /path/to/repo --dry-run # preview without writingFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.3. If `bun` is not installed: `curl -fsSL https://bun.sh/install | bash`Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.test_case "rm in /tmp" allow Bash "rm /tmp/test.txt"Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.await handleReadCommand('eval', ['/etc/passwd'], bm);Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.3. If `bun` is not installed: `curl -fsSL https://bun.sh/install | bash`Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.3. If `bun` is not installed: `curl -fsSL https://bun.sh/install | bash`Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.// Force kill if still aliveFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.expect(() => validateReadPath('/etc/passwd')).toThrow(/Absolute path must be within/);Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.Run full audit, then load previous \`design-baseline.json\`. Compare: per-category grade deltas, new findings, resolved findings. Output regression table in report.Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.expect(() => validateReadPath('src/../../etc/passwd')).toThrow(/Path traversal/);Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.expect(() => validateNavigationUrl('http://169.254.169.254/latest/meta-data/')).toThrow(/cloud metadata/i);Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.3. If `bun` is not installed: `curl -fsSL https://bun.sh/install | bash`Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.* Auto-update version/date lines in product docs before the release commit.Fix: Pin versions and sources; vendor anything fetched at install time.- `.license-guard.json` with WIP Computer defaultsFix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.// 3.75. Auto-update product docs version/dateFix: Pin versions and sources; vendor anything fetched at install time.console.log(`STALE: ${relOutput}`);Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.// 1. wip-readme-format /path/to/repo ... generates section filesFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.expect(() => validateReadPath('../../../etc/passwd')).toThrow(/Path traversal/);Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.expect(() => validateNavigationUrl('file:///etc/passwd')).toThrow(/scheme.*not allowed/i);Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.### B7. Update CLAUDE.mdFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.wip-readme-format /path/to/repo --check # validate, exit 0/1Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.// wip-readme-format /path/to/repo --check # validate existing README, exit 0/1Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.wip-readme-format /path/to/repo --dry-run # preview without writingFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.console.log(` [check] \u2705 ${output} — ${totalValid} commands, all valid`);Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.3. If `bun` is not installed: `curl -fsSL https://bun.sh/install | bash`Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.test('rm -rf /var/data warns with recursive delete message', () => {Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.console.log(formatGateOutput(safe, alerts));Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.describe('rm -rf / rm -r', () => {Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.Write DESIGN.md and CLAUDE.md (or update it) in the working directory.`,Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.const rawJson = '{"tool_input":{"command":\n"rm -rf /tmp/important"}}';Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.HOOK_CMD="npx @wipcomputer/wip-license-hook"Fix: Pin versions and sources; vendor anything fetched at install time.console.log(' Tip: LDM OS could not be installed automatically. Try: npm install -g @wipcomputer/wip-ldm-os');Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.console.log(' wip-repo-permissions check wipcomputer/wip-bridge');Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.1. **Toolbox repo** — dogfood first, manually write CLAUDE.mdFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.| `tools/wip-repos/check.mjs` | **Modify** — add CLAUDE.md validation |Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.… `guide/DEV-GUIDE.md` path that doesn't exist. Fix this in Step 6 as part of the CLAUDE.md update. This is a pre-existing bug (references `wip-dev-tools-private` which was rena…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.- `wip-file-guard` blocks Write on CLAUDE.md, blocks Edit removing >2 net linesFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.BUG 1: broken route — typo "prodcts" instead of "products"Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.judgeGoal: 'how to audit and update project documentation after code ships: README, ARCHITECTURE, CONTRIBUTING, CLAUDE.md, CHANGELOG, TODOS',Fix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.# Install: cp hooks/pre-pull.sh .git/hooks/pre-merge-commit && chmod +x .git/hooks/pre-merge-commitFix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.console.log(` \u274c ${output.padEnd(30)} — generated file missing! Run: bun run gen:skill-docs`);Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.- Onboarding-before-first-write: requires Read of `README.md`, `CLAUDE.md`, and any `*RUNBOOK*.md` / `*LANDMINES*.md` / `WORKFLOW*.md` at repo root before the first write.Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.1. Update `~/.openclaw/CLAUDE.md` (all three path references)Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.console.log(` \u26a0\ufe0f ${output.padEnd(30)} — no template`);Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.console.log(`Output has coverage summary: ${hasCoverage}`);Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.HOOK_CMD="npx @wipcomputer/wip-license-hook"Fix: Pin versions and sources; vendor anything fetched at install time.console.log(' --json Output detection results as JSON');Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.console.log(` [dry run] Would publish to npm (@wipcomputer scope)`);Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.console.log(` \u2705 ${tmpl.padEnd(30)} \u2192 ${output}`);Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.# Install: cp hooks/pre-push.sh .git/hooks/pre-push && chmod +x .git/hooks/pre-pushFix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.Added TOOLS.md, MEMORY.md, IDENTITY.md, SOUL.md, WHERE-TO-WRITE.md, HEARTBEAT.md to the shared state allowlist. Both agents can now write to workspace files on main without being blocked.Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.…m -rf .claude/skills/gstack/.git && cd .claude/skills/gstack && ./setup`** then add a "gstack" section to this project's CLAUDE.md that says to use the /browse skill from gstack for all web browsing, never use …Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.Added TOOLS.md, MEMORY.md, IDENTITY.md, SOUL.md, WHERE-TO-WRITE.md, HEARTBEAT.md to the shared state allowlist. Both agents can now write to workspace files on main without being blocked.Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.…ck.git ~/.claude/skills/gstack && cd ~/.claude/skills/gstack && ./setup`** then add a "gstack" section to CLAUDE.md that says to use the /browse skill from gstack for all web browsing, never use …Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.console.log(' wip-repo-permissions audit wipcomputer');Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.env: { ...process.env, ...env },Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.elif command -v npx &>/dev/null; thenFix: Pin versions and sources; vendor anything fetched at install time.describe('rm -rf / rm -r', () => {Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.| `tools/wip-repo-init/init.mjs` | **Modify** — add CLAUDE.md generation |Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.# Dev Update: CLAUDE.md Repo Paths Fix + Boot Sequence EnforcementFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.# Dev Update: CLAUDE.md Repo Paths Fix + Boot Sequence EnforcementFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.console.log(`Output has GAP markers: ${hasGap}`);Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.const { exitCode, output } = runHook(CAREFUL_SCRIPT, carefulInput('rm -rf /var/data'));Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.console.log(' wip-readme-format ... reformat a README to follow the WIP Computer standard');Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.console.log(`Output has TESTED markers: ${hasTested}`);Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.### B7. Update CLAUDE.mdFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.expect(() => validateNavigationUrl('http://192.168.1.1')).not.toThrow();Fix: Replace encoded or assembled strings with plain literals and remove undeclared network calls.Add to CLAUDE.md and the Dev Guide:Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.console.log(formatGateOutput(safe, alerts));Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.2. Update `~/.claude/CLAUDE.md` (mirror)Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.env: { ...process.env, ...env },Fix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.elif command -v npx &>/dev/null; thenFix: Pin versions and sources; vendor anything fetched at install time.'{"tool_name":"Edit","tool_input":{"file_path":"/foo/CLAUDE.md","old_string":"a\nb\nc\nd\ne\nf","new_string":"replaced"}}' \Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.echo '{"tool_name":"Write","tool_input":{"file_path":"/foo/CLAUDE.md"}}' | node guard.mjsFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.check "Block Write to CLAUDE.md" \Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.* Update the ecosystem section in a CLAUDE.md file.Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.'{"tool_name":"Write","tool_input":{"file_path":"/foo/CLAUDE.md","content":"new"}}' \Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.PASS: Block Write to CLAUDE.mdFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.check "Block Edit removing 5 lines from CLAUDE.md" \Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.PASS: Allow Edit adding lines to CLAUDE.mdFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.# Build a tmp repo with README + CLAUDE.md, then add a git worktree. TheFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.PASS: Block Edit removing 5 lines from CLAUDE.mdFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.console.log(` [update] ${repoName}/CLAUDE.md (${relevant.length} related repos)`);Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.git add README.md CLAUDE.mdFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.layer3_call "iso-b" "Write" "$LAYER3_REPO/iso-b1.md" "deny" "iso: session B still needs its own CLAUDE.md read"Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.check "Allow Edit adding lines to CLAUDE.md" \Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.test_case "Write auto-memory index" allow Write "$HOME/.claude/projects/-Users-lesa-example/memory/MEMORY.md"Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.// Required reads for first write in a repo. Root-level README.md, CLAUDE.md,Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.git add README.md CLAUDE.mdFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.Cost to the agent
Tokens estimated at four characters each. The tool list is re-read by the agent on every turn, so it is the number to compare between servers that do the same job: ranked across the index.
Known vulnerabilities
Vulnerability advisories (OSV) cover npm and PyPI packages; this registry has no advisory database.
What would make this Verified
16 points to 80 · weights: behavior 45%, static 25%, prompt 15%, publisher 15%- Destructive command against a filesystem root or device≈ +9 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
- Unclaimed publisher≈ +8 on the totalClaim parkertoddbrooks: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
- Hidden instructions≈ +6 on the totalRemove HTML comments, zero-width characters and white-on-white text that carry instructions. Everything the agent is meant to read should be visible to the person reading the file.
- Pipe-to-shell pattern≈ +4 on the totalBuild commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
- Destructive file system command detected≈ +4 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
- Auto-update or dynamic fetch configuration≈ +4 on the totalPin versions and sources; vendor anything fetched at install time.