Borrowing it
Nothing to install: this file belongs to SebaBoler/vanguard. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/SebaBoler/vanguard/main/CLAUDE.mdgit clone --depth 1 https://github.com/SebaBoler/vanguardWrote 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/instructions/sebaboler/vanguard/claude-md)<a href="https://agentmods.dev/instructions/sebaboler/vanguard/claude-md"><img src="https://agentmods.dev/badge/instructions/sebaboler/vanguard/claude-md/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/instructions/sebaboler/vanguard/claude-md"><img src="https://agentmods.dev/badge/instructions/sebaboler/vanguard/claude-md.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 | $0.00559 | $0.00559 |
| Opus 5.5 | $0.00224 | $0.00224 |
| Sonnet 5.5 | $0.00112 | $0.00112 |
| Haiku 4.5 | $0.00056 | $0.00056 |
Grade B, and why
vanguard CLAUDE.md 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 22d 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.
Unrestricted tool accessmediumExcessive agency
A wildcard tool grant or "run any command" leaves no least-privilege boundary at all.
- **Never modify CI config: `.github/workflows/`, `.github/actions/`, `.gitlab-ci.yml`, `.gitlab/**.yml`** (authoritative check: `workflowPaths()` in `src/core/vanguard.ts`, which also covers `.gitlab/**/*.yaml` at any d How it starts
The opening of the file, as written. The whole thing — 22 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
Vanguard is a standalone TypeScript framework — an autonomous software factory. Strict TypeScript, ES modules with explicit .js import extensions, Node 24+. Tests are co-located as *.test.ts (Vitest).
Hard constraints
- Never modify CI config:
.github/workflows/,.github/actions/,.gitlab-ci.yml,.gitlab/**.yml(authoritative check:workflowPaths()insrc/core/vanguard.ts, which also covers.gitlab/**/*.yamlat any depth and files standing in for those directories). This is a security boundary: an agent that can commit CI config can run arbitrary code with the repo's secrets on the next event. It is also enforced mechanically — copy-back drops those paths, and the CIGITHUB_TOKENcannot push workflow changes, so such a push is rejected and the run's work is lost. If a task seems to need a CI-config change, make the code/doc change instead and describe the needed edit in the PR body for a human to apply by hand. - Run
pnpm typecheckandpnpm testbefore signalling completion.
Style
- Match the surrounding code: comment density, naming, idiom. Keep diffs minimal.
- Explicit return types; prefer
const; early returns; functional where it fits.
Review guidance
For the PR reviewer. The CI-config constraint above binds the autonomous run (implement, revise); humans change CI config through PRs, which are reviewed, not refused.
- A CI-config edit is not a finding for the constraint itself when the API-sourced PR author (never a name claimed in the PR text, commits or diff) passes the author gate in
.github/workflows/vanguard-pr-review.yml(bots do not; Dependabot's action-pin bumps never reach this reviewer). Review the content as the highest-risk part of the diff: trigger type (pull_request_target),permissions:, actor gates, secret exposure, no${{ }}interpolation insiderun:. A CI-config edit by any other author is itself a finding. - The
workflow-scoped token a maintainer pushes with is personal and local, never a CI secret (VANGUARD_PUSH_TOKENmust not carryworkflowscope). - When every actionable item is on a CI-config path, say so in the verdict and do not recommend
needs revision: the reviser drops edits to those paths and ends in "no changes — skipping push" without posting, leavingneeds revisionandvanguard:revisingon the PR. With mixed feedback it fixes the rest and discloses the dropped paths.
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.
- 22d ago Changed · +8 lines · +359 tokens per session scan A → B 3b6c36ad8c7e
- 2mo ago First seen · 14 lines · 200 tokens per session scan A 8b7535dd0ed5
vanguard CLAUDE.md is an instructions file published in the GitHub repository SebaBoler/vanguard (4 stars, last pushed 9d ago), licensed MIT. It adds 559 tokens to every session, about $0.0022 per session on Opus 5.5. A static security scan graded it B with 1 finding (unrestricted tool access). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, cache eligibility and rollout and applying the notice (cutover).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode copilot-instructions.md
Copilot instructions for microsoft/vscode, covering vs code copilot instructions, project overview, root folders, core architecture (src/ folder) and built-in extensions (extensions/ folder).
deepseek-harness AGENTS.md
AGENTS.md instructions for deepseek-ai/deepseek-harness, covering agents.md, pre-stable apis and released session data, repository layout, commands and host sandbox failures.