Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/vscarpenter/gsd-task-manager/agents-mdgit clone --depth 1 https://github.com/vscarpenter/gsd-task-managerWhat 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 | $0.01014 | $0.01014 |
| Opus 5 | $0.00507 | $0.00507 |
| Sonnet 5 | $0.00203 | $0.00203 |
| Haiku 4.5 | $0.00101 | $0.00101 |
Grade A, and why
gsd-task-manager AGENTS.md scanned grade A with 0 findings 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 3d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 74 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Operating notes for AI agents working in this repository. General project
guidance, architecture, and the canonical command list live in CLAUDE.md and
coding-standards.md — read those first. This file only adds Cursor Cloud
environment caveats.
Cursor Cloud specific instructions
Runtime / package manager. This project uses Bun (lockfile bun.lock,
Bun workspaces include packages/*). Bun is installed by the startup update
script (it is not on the base image) and is on PATH via ~/.bashrc. The
standard dev/test/build commands are documented in CLAUDE.md (bun dev,
bun run test, bun typecheck, bun lint, bun run build). Use bun run test
(not bun test, which invokes Bun's own runner instead of Vitest).
Dev server. bun dev serves at http://localhost:3000. The app is 100%
client-side (every component is "use client"; no SSR/API routes; static
export). On first load the root URL redirects to /about (onboarding); enter the
matrix via the onboarding "Open App" CTA. All data lives in the browser's
IndexedDB (Dexie), so a fresh browser profile starts empty — there is no backend
to seed for basic task CRUD.
Optional backend (PocketBase) is NOT required for local dev. Cloud sync,
realtime SSE, OAuth, and the packages/mcp-server workspace only matter when a
PocketBase instance is reachable (NEXT_PUBLIC_POCKETBASE_URL, defaults to
http://127.0.0.1:8090 on localhost). The core product runs and is testable
fully local-only without it.
Lint and TypeScript compatibility. bun lint is expected to run. The root
workspace deliberately installs the TypeScript 7 CLI as @typescript/native
while keeping [email protected] for the compiler API. The bun typecheck
script explicitly invokes the native alias; Next.js and typescript-eslint
continue to import TypeScript 6. Keep both packages until those tools support
the new TypeScript API. ESLint 10 also relies on the explicit React version in
eslint.config.mjs because [email protected] still calls an API
ESLint 10 removed. The standalone MCP workspace has no compiler-API consumer
and uses TypeScript 7 directly. A blanket bun update --latest will try to
replace the root TypeScript 6 API package; use the interactive updater and leave
that package pinned. The build-config regression test enforces this split.
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.
- 3d ago First seen · 74 lines · 1,014 tokens per session scan A f187082b26a2
gsd-task-manager AGENTS.md is an instructions file published in the GitHub repository vscarpenter/gsd-task-manager (24 stars, last pushed 4d ago), licensed MIT. It adds 1,014 tokens to every session, about $0.0051 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
mediawork-mcp AGENTS.md
AGENTS.md instructions for screen-arts/mediawork-mcp, covering mcp.mediawork.io for agents, the two rules that define this project, layout, conventions and running it.
lastest CLAUDE.md
Claude Code instructions for las-team/lastest, covering claude.md, critical rules, commands, tesseract. part of the default compose stack; set and ocrserviceurl=http://localhost:8891 in .env.local.
lastest AGENTS.md
AGENTS.md instructions for las-team/lastest, covering agents.md — lastest, what is lastest?, mcp server, available tools (50 total, all prefixed lastest) and typical workflow.
sp500-mcp-server AGENTS.md
AGENTS.md instructions for zhensherlock/sp500-mcp-server, covering agents.md - sp500-mcp-server, repo shape, commands, runtime and test gotchas and mcp tool wiring.
sp500-mcp-server CLAUDE.md
Claude Code instructions for zhensherlock/sp500-mcp-server: Use AGENTS.md as the primary project instruction file.
CountryStateCity CLAUDE.md
Claude Code instructions for tansuasici/CountryStateCity, covering country state city - project guide, project overview, tech stack, project structure and key commands.