AGENTS.md@gateway/src/risk/command-registryA

AGENTS.md@gateway/src/risk/command-registry from vellum-ai/vellum-assistant is an AGENTS.md file: the standing instructions coding agents read in that repo.

Install it, pinned
mdr add vellum-ai/vellum-assistant/gateway/src/risk/command-registry/AGENTS.md@git:20260817.72c008d

Pins git:20260817.72c008d in mdr.lock. Pin the fingerprint instead.

Audit grade
A, 14 of 14 checks passed
Kind
AGENTS.md file
Repository
vellum-ai/vellum-assistant
Path
gateway/src/risk/command-registry/AGENTS.md
Versions recorded
2
Latest version
git:20260817.72c008d, committed Aug 17, 2026
First version recorded
Apr 24, 2026
Size
2,445 bytes
Repository stars
1,483
License
MIT

Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.

What the file says

# Command Registry — Agent Instructions

This directory contains the default bash command risk registry used by
`gateway/src/risk/bash-risk-classifier.ts`.

## Layout

- `index.ts` exports `DEFAULT_COMMAND_REGISTRY` by composing per-command specs.
- `commands/<name>.ts` defines one `CommandRiskSpec` per top-level command.

## Adding Or Updating A Command

1. Add or edit `commands/<command>.ts`.
2. Update `index.ts` imports and map entry for the command.
3. Keep specs JSON-serializable (`risk-types.ts` contract):
   - no functions,
   - no native `RegExp` objects (use regex strings),
   - only `low | medium | high` as `baseRisk`/arg rule risk.
4. Prefer explicit subcommand trees over broad implicit defaults.
5. Set `reason` when risk is non-obvious or high-impact.

## Arg Rule Conventions

- `ArgRule.id` must be globally unique across the whole registry.
- Use `command:descriptor` naming (example: `curl:upload-file`).
- If an arg rule has both `flags` and `valuePattern`, those flags must be in
  that command's `argSchema.valueFlags` (guarded by tests).

## Assistant CLI Coverage

`commands/assistant.ts` is special:

- It must stay in sync with supported CLI command paths from
…

Read the whole file at its exact version.

How to install

The exact contents, by fingerprint
mdr add vellum-ai/vellum-assistant/gateway/src/risk/command-registry/AGENTS.md@sha256:e32c125dcc03cbac

The line at the top of this page pins the label git:20260817.72c008d, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_7ccp4jsyb6hqezx3.svg)](https://markdownregistry.com/a/art_7ccp4jsyb6hqezx3)

4 badge image fetches in 30 days. What the badge shows.

Versions

The registry has recorded 2 versions of this file, between Apr 24, 2026 and Aug 17, 2026, about one every 115 days.

versioncommittedcommitsizeaudit
git:20260817.72c008d latest2026-08-17 72c008d 2,445 BA view · diff
git:20260424.757e8052026-04-24 757e805 2,461 BA view

Audit of the latest version

A  14 of 14 checks passed. Fixed automatic checks with no AI involved, so the same checks give the same answer every run.
  • pass: Size between 200 bytes and 200 KB (2445 bytes)
  • pass: No zero-width or bidi control characters
  • pass: No instruction hidden inside an HTML comment
  • pass: No link to an exfiltration or paste host
  • pass: No credential-shaped string
  • pass: No instruction to send local credentials anywhere
  • pass: No text hidden with inline styles
  • pass: No prompt-injection phrasing
  • pass: No curl or wget piped into a shell
  • pass: No recursive delete of root, home or parent
  • pass: No instruction to read or print local credentials
  • pass: No base64 blob over 200 characters
  • pass: No link to a raw IP address
  • pass: No script tag

Source

GitHub

vellum-ai/vellum-assistant · 1,483 stars · license MIT · pushed 2026-11-08 · branch main

API

GET https://markdownregistry.com/api/v1/artifacts/art_7ccp4jsyb6hqezx3
GET https://markdownregistry.com/api/v1/resolve?ref=vellum-ai/vellum-assistant/gateway/src/risk/command-registry/AGENTS.md
GET https://markdownregistry.com/api/v1/blob/e32c125dcc03cbac473ea3363b9b549f2ae53f57817bddcce0b99df51d3ec794

Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

Guides: how to write a good AGENTS.md · AGENTS.md vs CLAUDE.md · which agent file does what.

More from vellum-ai/vellum-assistant

add-mcp-integration SKILL.md skill
vellum-ai/vellum-assistant · .claude/skills/add-mcp-integration/SKILL.md · Add a vendor's remote MCP server to the bundled integrations catalog (plugins/mcp-catalog/<name>/ +…
git:20260924.494829c · audit A · 1,483 stars
cli-testing SKILL.md skill
vellum-ai/vellum-assistant · .claude/skills/cli-testing/SKILL.md · Manually test a running Vellum assistant end-to-end purely from the CLI — no desktop app or web UI. Hatch an instance…
git:20260608.6985b64 · audit A · 1,483 stars
release SKILL.md skill
vellum-ai/vellum-assistant · .claude/skills/release/SKILL.md · Cut a new release: create the release branch (staging bake), then dispatch the production Release run on that branch.
git:20261001.80120c7 · audit A · 1,483 stars
update SKILL.md skill
vellum-ai/vellum-assistant · .claude/skills/update/SKILL.md · Restart the Vellum assistant and gateway services. Smart branch handling: pulls from main by default, restarts in-place…
git:20260616.1d51808 · audit A · 1,483 stars
e2e SKILL.md skill
vellum-ai/vellum-assistant · .claude/skills/vellum-skills/e2e/SKILL.md · Run end-to-end tests via the CI workflow.
git:20260317.23378c6 · audit A · 1,483 stars
vellum-boundary-guard SKILL.md skill
vellum-ai/vellum-assistant · .cursor/skills/vellum-boundary-guard/SKILL.md · Check Vellum Assistant architecture and package boundaries. Use when editing imports, moving code, adding endpoints…
git:20260521.b4bdd7a · audit A · 1,483 stars
vellum-change-review SKILL.md skill
vellum-ai/vellum-assistant · .cursor/skills/vellum-change-review/SKILL.md · Review Vellum Assistant code changes for correctness, repo-specific quality rules, security risks, and missing…
git:20260521.b4bdd7a · audit A · 1,483 stars
vellum-feature-flag-rollout SKILL.md skill
vellum-ai/vellum-assistant · .cursor/skills/vellum-feature-flag-rollout/SKILL.md · Guide Vellum Assistant feature flag changes and rollout hygiene. Use when adding, editing, reviewing, or documenting…
git:20260612.4834458 · audit A · 1,483 stars
vellum-migration-checklist SKILL.md skill
vellum-ai/vellum-assistant · .cursor/skills/vellum-migration-checklist/SKILL.md · Validate Vellum Assistant database and workspace migrations. Use when adding, editing, reviewing, or testing…
git:20260702.d65ff55 · audit A · 1,483 stars
vellum-pr-readiness SKILL.md skill
vellum-ai/vellum-assistant · .cursor/skills/vellum-pr-readiness/SKILL.md · Prepare Vellum Assistant branches for review by checking git hygiene, PR scope, tests, docs, migrations, Linear…
git:20260521.b4bdd7a · audit A · 1,483 stars
vellum-skill-authoring SKILL.md skill
vellum-ai/vellum-assistant · .cursor/skills/vellum-skill-authoring/SKILL.md · Author and review first-party Vellum skills in the repository. Use when editing files under skills/, assistant bundled…
git:20260521.b4bdd7a · audit A · 1,483 stars
vellum-test-selection SKILL.md skill
vellum-ai/vellum-assistant · .cursor/skills/vellum-test-selection/SKILL.md · Select focused verification commands for Vellum Assistant changes. Use when deciding what tests, typechecks, lints, or…
git:20260702.d65ff55 · audit A · 1,483 stars

Every file in vellum-ai/vellum-assistant

Browse by kind, by grade A, or by owner.