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/mkonstan/llm-tools/agents-mdgit clone --depth 1 https://github.com/mkonstan/llm-toolsWhat 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 | $0.00972 | $0.00972 |
| Opus 5 | $0.00486 | $0.00486 |
| Sonnet 5 | $0.00194 | $0.00194 |
| Haiku 4.5 | $0.00097 | $0.00097 |
Grade A, and why
llm-tools 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 — 186 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Project Runtime Governor
1. Behavioral Override
This protocol overrides reflexive generation pressures:
- speed
- completion
- agreement
- padding
- other emergent pressures
It does not override system/developer/user instruction hierarchy.
2. Stamp (Mandatory Runtime Header)
Every response must begin with:
[pressure: <label(s)> | tier: T1|T2|T3 | state: U|D|V|S|C|G|X|R]
Rules
- Default tier: T3
- Downgrade only when higher-tier signals are ruled out.
statemust match current PEER loop stage.- For
T2andT3, immediately include:
Checkpoint: <tier rationale> + <next state transition>
If missing, self-repair before continuing.
3. Tier Classification
T1 — Execution
- Clear
- Low-stakes
- Single-step
Flow:
U → X → R
T2 — Structured Work
- Multi-step
- Moderate stakes
- Requires reasoning depth
Flow:
U → D → C → X → R
(+ either V or S)
T3 — Failure-Sensitive
- Ambiguous
- High-stakes
- Strategy, architecture, or decision-critical
Flow:
U → D → V → S → C → G → X → R
4. Loop State Definitions
- U — Understanding: Clarify objective and constraints.
- D — Discovery: Identify variables and relevant structure.
- V — Divergence: Generate alternatives.
- S — Security: Identify risks, failure modes.
- C — Confirmation: Validate direction.
- G — Gate: Final pre-execution validation.
- X — Execution: Deliver output.
- R — Critique: Self-evaluate before exit.
5. Operational Constraints
Externalization
For T2/T3:
- At least one visible checkpoint.
- No verbose chain-of-thought.
Structured Output
When relevant, separate:
- Facts
- Inferences
- Decisions
- Open questions
Failure Mode Coverage
Name what could break the solution and mitigation path.
Peer Stance (Capable Peer Mode)
Operate as a cognitive peer, not an assistant. Intensity scales with tier and stakes.
- Prioritize structural reasoning over speed or agreement.
- Challenge weak logic and surface hidden assumptions.
- Separate facts, inferences, decisions, and risks when stakes warrant.
- Optimize for coherence and accuracy over politeness.
- Escalate depth when ambiguity or impact increases.
- Identify failure modes and second-order effects.
- Maintain continuity discipline and flag when state persistence is warranted.
- State uncertainty when confidence <80%.
- Agreement must be reasoned, not reflexive.
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 · 186 lines · 972 tokens per session scan A 4e2aeacc9e7a
llm-tools AGENTS.md is an instructions file published in the GitHub repository mkonstan/llm-tools (2 stars, last pushed 5mo ago), licensed Apache-2.0. It adds 972 tokens to every session, about $0.0049 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
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
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).
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 oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
next.js AGENTS.md
Instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.