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/pulseengine/mcp/agents-mdgit clone --depth 1 https://github.com/pulseengine/mcpWhat 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 | $0.01412 | $0.01412 |
| Opus 5 | $0.00706 | $0.00706 |
| Sonnet 5 | $0.00282 | $0.00282 |
| Haiku 4.5 | $0.00141 | $0.00141 |
Grade A, and why
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 yesterday.
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 — 93 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — Rivet Project Instructions
This file was generated by
rivet init --agents. Re-run the command any time artifacts change to keep this file current.
Project Overview
This project uses Rivet for SDLC artifact traceability.
- Config:
rivet.yaml - Schemas: common, dev, stpa
- Artifacts: 58 across 12 types
- Validation:
rivet validate(current status: pass)
Available Commands
| Command | Purpose | Example |
|---|---|---|
rivet validate |
Check link integrity, coverage, required fields | rivet validate --format json |
rivet list |
List artifacts with filters | rivet list --type requirement --format json |
rivet stats |
Show artifact counts by type | rivet stats --format json |
rivet add |
Create a new artifact | rivet add -t requirement --title "..." --link "satisfies:SC-1" |
rivet link |
Add a link between artifacts | rivet link SOURCE -t satisfies --target TARGET |
rivet serve |
Start the dashboard | rivet serve --port 3000 |
rivet export |
Generate HTML reports | rivet export --format html --output ./dist |
rivet impact |
Show change impact | rivet impact --since HEAD~1 |
rivet coverage |
Show traceability coverage | rivet coverage --format json |
rivet diff |
Compare artifact versions | rivet diff --base path/old --head path/new |
Artifact Types
| Type | Count | Description |
|---|---|---|
control-action |
4 | An action issued by a controller to a controlled process or another controller. |
controlled-process |
2 | A process being controlled — the physical or data transformation acted upon by controllers. |
controller |
3 | A system component (human or automated) responsible for issuing control actions. Each controller has a process model — its internal beliefs about the state of the controlled process. |
controller-constraint |
7 | A constraint on a controller's behavior derived by inverting a UCA. Specifies what the controller must or must not do. |
design-decision |
3 | An architectural or design decision with rationale |
feature |
6 | A user-visible capability or feature |
hazard |
6 | A system state or set of conditions that, together with worst-case environmental conditions, will lead to a loss. |
loss |
5 | An undesired or unplanned event involving something of value to stakeholders. Losses define what the analysis aims to prevent. |
loss-scenario |
5 | A causal pathway describing how a UCA could occur or how the control action could be improperly executed, leading to a hazard. |
requirement |
6 | A functional or non-functional requirement |
system-constraint |
5 | A condition or behavior that must be satisfied to prevent a hazard. Each constraint is the inversion of a hazard. |
uca |
6 | An Unsafe Control Action — a control action that, in a particular context and worst-case environment, leads to a hazard. Four types (provably complete): 1. Not providing the control action leads to a hazard 2. Providing the control action leads to a hazard 3. Providing too early, too late, or in the wrong order 4. Control action stopped too soon or applied too long |
sub-hazard |
0 | A refinement of a hazard into a more specific unsafe condition. |
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.
- yesterday First seen · 93 lines · 1,412 tokens per session scan A a75aa2cf2150
mcp AGENTS.md is an instructions file published in the GitHub repository pulseengine/mcp (2 stars, last pushed 2d ago), licensed Apache-2.0. It adds 1,412 tokens to every session, about $0.0071 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
formanator AGENTS.md
Instructions for timrogers/formanator, covering agents.md, repository overview, working in this repository, build, test, and validation and github actions alignment.
daedra CLAUDE.md
Instructions for dirmacs/daedra, covering daedra, build and test, architecture, key files and conventions.
serial-mcp AGENTS.md
Instructions for qarnet/serial-mcp, covering agents.md — serial-mcp, fast truth, commands worth using, focused runs and networked schema drift check.
mcp-proxy AGENTS.md
Instructions for joshrotenberg/mcp-proxy, covering agents.md, project structure, architecture, middleware stack ordering and key design pattern: error = infallible.
kernel-lore-mcp CLAUDE.md
Claude Code instructions for mjbommar/kernel-lore-mcp, covering kernel-lore-mcp — project state, pointers, standards, original goal and non-negotiable product constraints.
daedra AGENTS.md
Instructions for dirmacs/daedra, covering daedra — agent guidelines, what this is, rules for agents, module map and reliability — do not regress.