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/stevekinney/temporal-mcp/agents-mdgit clone --depth 1 https://github.com/stevekinney/temporal-mcpWrote 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/stevekinney/temporal-mcp/agents-md)<a href="https://agentmods.dev/instructions/stevekinney/temporal-mcp/agents-md"><img src="https://agentmods.dev/badge/instructions/stevekinney/temporal-mcp/agents-md.svg" alt="Measured on agentmods" 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 | $0.01359 | $0.01359 |
| Opus 5 | $0.00679 | $0.00679 |
| Sonnet 5 | $0.00272 | $0.00272 |
| Haiku 4.5 | $0.00136 | $0.00136 |
Grade A, and why
temporal-mcp AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
Copies of this mod
1 near-identical copy found in the catalogue:
- temporal-mcp CLAUDE.md — 100% identical, 0 lines differ
How it starts
The opening of the file, as written. The whole thing — 145 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Default to using Bun instead of Node.js.
- Use
bun <file>instead ofnode <file>orts-node <file> - Use
bun testinstead ofjestorvitest - Use
bun build <file.html|file.ts|file.css>instead ofwebpackoresbuild - Use
bun installinstead ofnpm installoryarn installorpnpm install - Use
bun run <script>instead ofnpm run <script>oryarn run <script>orpnpm run <script> - Use
bunx <package> <command>instead ofnpx <package> <command> - Bun automatically loads .env, so don't use dotenv.
APIs
Bun.serve()supports WebSockets, HTTPS, and routes. Don't useexpress.bun:sqlitefor SQLite. Don't usebetter-sqlite3.Bun.redisfor Redis. Don't useioredis.Bun.sqlfor Postgres. Don't usepgorpostgres.js.WebSocketis built-in. Don't usews.- Prefer
Bun.fileovernode:fs's readFile/writeFile - Bun.$
lsinstead of execa.
Path Containment Safety
- When containment checks are based on
path.relative, treat absolute outputs (including Windows drive-letter forms) as outside the root boundary. - Normalize path separators before prefix checks so containment logic remains separator-agnostic across platforms.
Review Memory Invariants
- Resolve policy scope first, then use the resolved profile/namespace consistently for policy evaluation, audit context, and client acquisition.
- Evaluate Temporal policy against resolved profile/namespace values, not raw optional inputs.
- Enforce policy documentation and behavior that
allowPatternsanddenyPatternsrun before mode checks in every mode. - Treat policy allowlist misses as a distinct decision code from deny-pattern matches.
- Keep capability-matrix contract lookup fail-closed (
TOOL_NOT_FOUND) and centralized in shared helpers such asrequireToolContract. - Never skip policy evaluation when a tool/resource contract lookup fails.
- Resources must enforce the same policy and redaction guarantees as equivalent tools.
- Resource resolvers must emit audit
tool_call,policy_decision, andtool_resultevents with request context. - Build request context exactly once per handler, then mutate
requestContext.profileafter scope resolution. - Emit exactly one
tool_callbefore anytool_result; if early scope resolution can throw, add catch-path fallbacktool_calllogging. - Emit
tool_resultlogs on all early-error returns so failed paths remain observable. - Reuse the original request context through helper-based policy checks to preserve request IDs.
- Keep audit argument-redaction and response-redaction pattern sets aligned.
- For unsafe/break-glass gating, check environment-variable presence (
=== undefined) rather than truthiness. - Documentation path validation must include realpath boundary checks to block symlink escapes.
- Keep path-containment logic separator-agnostic and avoid redundant/dead traversal guards once normalized boundary checks exist.
- Status endpoints that summarize index data must rely on metadata persisted at index-build time.
- Any refresh flow that claims search readiness must rebuild and persist the index in the same operation.
- Recovery after failed repository sync/subprocess steps must validate exit codes and return structured errors.
- Decode subprocess
Uint8Arrayoutput explicitly withTextDecoderbefore parsing stderr/stdout-derived metadata. - SSRF guards for codec endpoints must normalize IPv6 hostnames and treat IPv6-mapped IPv4 addresses as the underlying IPv4 for private-range checks.
- Codec decode paths must treat any defined payload (including falsy values) as valid; only
undefinedis an empty result. - Legacy registration entrypoints must delegate to the shared
ToolRegistrationContextpipeline to preserve policy/audit/redaction behavior. - Package export maps must reference real entrypoint files; maintain the
src/index.tsbarrel when exporting".". - Legacy registration tests should verify stable workflow-level behavior and mock policy-scope connection methods used by shared registration.
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.
- 4d ago First seen · 145 lines · 1,359 tokens per session scan A 5ae1433e1e43
temporal-mcp AGENTS.md is an instructions file published in the GitHub repository stevekinney/temporal-mcp (5 stars, last pushed 7d ago), licensed MIT. It adds 1,359 tokens to every session, about $0.0068 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 2026-08-31.
Other instructions, from other repositories
vscode design-philosophy.instructions.md
VS Code design philosophy — the shared Values→Principles→Moves vocabulary for reasoning about UI in design terms rather than raw pixels. Use when creating, editing, or reviewing any visual surface (CSS, DOM, theming, icons, motion). Name the value/principle before reaching for a token.
deepagents AGENTS.md
AGENTS.md instructions for langchain-ai/deepagents, covering global development guidelines for the deep agents monorepo, corridor security analysis, development workflow, suppressing ruff rules and pr conventions.
walkerOS CLAUDE.md
Claude Code instructions for elbwalker/walkerOS, covering claude.md and important: use agent.md.
solo copilot-instructions.md
Copilot instructions for hiero-ledger/solo, covering github copilot instructions for hiero-ledger/solo and highest-frequency rules.
openmcp-client CLAUDE.md
Instructions for LSTM-Kirigaya/openmcp-client, covering claude.md, development commands, setup and installation, development and service development.
Kokoro-Engine CLAUDE.md
Instructions for chyinan/Kokoro-Engine, covering claude.md, git 规范, 项目概述, 常用命令 and 安装依赖.