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/mikehostetler/wreckit/agents-mdgit clone --depth 1 https://github.com/mikehostetler/wreckitWrote 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/mikehostetler/wreckit/agents-md)<a href="https://agentmods.dev/instructions/mikehostetler/wreckit/agents-md"><img src="https://agentmods.dev/badge/instructions/mikehostetler/wreckit/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 | $0.02778 | $0.02778 |
| Opus 5 | $0.01389 | $0.01389 |
| Sonnet 5 | $0.00556 | $0.00556 |
| Haiku 4.5 | $0.00278 | $0.00278 |
Grade B, and why
wreckit AGENTS.md scanned grade B with 1 finding 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
4. `~/.claude/settings.json` `env` (Claude user settings) How it starts
The opening of the file, as written. The whole thing — 295 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Wreckit Agent Guidelines
Specifications
IMPORTANT: Before implementing any feature, consult the specifications in specs/README.md.
- Assume NOT implemented. Many specs describe planned features that may not yet exist in the codebase.
- Check the codebase first. Before concluding something is or isn't implemented, search the actual code. Specs describe intent; code describes reality.
- Use specs as guidance. When implementing a feature, follow the design patterns, types, and architecture defined in the relevant spec.
- Spec index:
specs/README.mdlists all specifications organized by phase.
Commands
Building & Testing
| Command | Does |
|---|---|
bun build |
Build the CLI |
bun test |
Run all tests |
bun test src/__tests__/foo.test.ts |
Run single test file |
bun run typecheck |
Type check the codebase |
bun run lint |
Lint the codebase |
CLI Commands
| Command | Does |
|---|---|
wreckit |
Run all incomplete items (research → plan → implement → PR) |
wreckit next |
Run next incomplete item |
wreckit run <id> |
Run single item through all phases (id: 1, 2, or 001-slug) |
wreckit ideas < FILE |
Ingest ideas (create idea items) |
wreckit learn |
Extract patterns and compile into reusable skills |
wreckit status |
List all items + state |
wreckit list |
List items (with optional --state filtering) |
wreckit show <id> |
Show item details |
wreckit init |
Initialize .wreckit/ in repo |
wreckit doctor |
Validate items, fix broken state |
wreckit rollback <id> |
Rollback a direct-merge item to pre-merge state |
wreckit geneticist |
Recursive prompt self-optimization (evolution) |
Phase Commands (debugging)
| Command | Transition |
|---|---|
wreckit research <id> |
idea → researched |
wreckit plan <id> |
researched → planned |
wreckit implement <id> |
planned → implementing |
wreckit pr <id> |
implementing → in_pr |
wreckit complete <id> |
in_pr → done |
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 · 295 lines · 2,778 tokens per session scan B c490a6b4a8b1
wreckit AGENTS.md is an instructions file published in the GitHub repository mikehostetler/wreckit (130 stars, last pushed 4mo ago), licensed MIT. It adds 2,778 tokens to every session, about $0.0139 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
ESAA-Core AGENTS.md
Instructions for elzobrito/ESAA-Core, covering agents.md — contrato operacional codex/esaa, 1. autoridade e termos, 2. cli e runner, ou configure o runner uma vez por sessão and 3. concorrência.
Kanvas AGENTS.md
Instructions for XMihura/Kanvas, covering canvas workflow — agent instructions, critical rule, session protocol, 1. start of session — read the board and 2. pick a task.
delivery-loop CLAUDE.md
Instructions for blakemartz/delivery-loop, covering claude.md, what this is, the cardinal rule: everything must stay repo-agnostic, layout and config: the one seam.
taskcenter AGENTS.md
AGENTS.md instructions for xiaogezi/taskcenter, covering taskcenter agent rules, safety boundaries, required workflow and verification.
idd-skill idd-resume.instructions.md
Instructions for kurone-kito/idd-skill, covering idd — resume phase, required inputs, step 0 — route classifier, operator-present release and step 1 — identify claim state.
codecrucible AGENTS.md
Instructions for block/codecrucible, covering agent instructions, quick reference and landing the plane (session completion).