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/shafthq/shaft_engine/agents-mdgit clone --depth 1 https://github.com/ShaftHQ/SHAFT_ENGINEWhat 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 | $0.00727 | $0.00727 |
| Opus 5 | $0.00364 | $0.00364 |
| Sonnet 5 | $0.00145 | $0.00145 |
| Haiku 4.5 | $0.00073 | $0.00073 |
Grade A, and why
SHAFT_ENGINE 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 — 56 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Repository facts
ChaosEngine by Mohab Mohie. SHAFT_ENGINE is a Maven Java automation framework:
core shaft-engine/, optional shaft-*, IntelliJ plugin shaft-intellij/,
and CI tooling under scripts/ci/. Configuration wins. Start from the
requested goal and affected files.
Canonical policy
Before every task, read and follow ChaosEngine. It is the only router and working-policy owner. Its selected profile and references own lifecycle hooks, roles, capability levels, skills, tools, Caveman, Ponytail, TDD, research, memory, task isolation, cleanup, delegation, review, delivery, and completion. Do not restate those policies in host adapters.
Cleanup scope is defined only by
cleanup-scopes. Harness duty
ownership is machine-readable in
agent_ownership.json. CLAUDE.md, host
skills, and provider guidance remain thin pointers to the canonical router.
Repository safety
- Read live files first, preserve unrelated and pre-existing work, and keep changes within the requested repository and task scope.
- Do not launch GUI applications, browsers, editors, installers, servers, or watchers without explicit authorization.
- Keep Maven tests scoped and headless with
-Dallure.automaticallyOpen=false -DheadlessExecution=true. - Never track generated reports, binaries, caches,
target/, build output, Graphify output, MemPalace runtime indexes, secrets, or machine-local state. - Preserve public APIs; reproduce defects with focused regressions. Functional documentation changes remain a separate PR and use the configured docs root.
- Validate harness changes with
py -3 scripts/ci/validate_agent_setup.py --skip-externaland the smallest directly affected tests. Inspect result artifacts rather than banners alone.
Before every task, follow the canonical ChaosEngine. Use .chaos-engine/tool.py for the project-local Memory, MemPalace, and Graphify tools.
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 · 56 lines · 727 tokens per session scan A 30b5789b925c
SHAFT_ENGINE AGENTS.md is an instructions file published in the GitHub repository ShaftHQ/SHAFT_ENGINE (407 stars, last pushed yesterday), licensed MIT. It adds 727 tokens to every session, about $0.0036 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
intervals-icu-mcp CLAUDE.md
Instructions for hhopke/intervals-icu-mcp, covering claude.md, project overview, development commands, architecture (quick reference) and tool categories.
flyto-core CLAUDE.md
Instructions for flytohub/flyto-core, covering claude notes, cross-agent handoff and shared code intelligence.
rosetta AGENTS.md
Instructions for tikoci/rosetta, covering codex instructions for rosetta, first reads, development defaults and mcp and client config.
pharo-smalltalk-interop-mcp-server CLAUDE.md
Instructions for mumez/pharo-smalltalk-interop-mcp-server, covering claude.md, project overview, development setup, environment variables and common commands.
rosetta llm-instruction-files-excluded-from-mdlint.instructions.md
LLM-targeted instruction files are excluded from markdownlint so authoring rules for human docs do not distort agent-facing guidance.
mcp-server-powerpoint code-review.instructions.md
Instructions for sbroenne/mcp-server-powerpoint, covering copilot code review, interop and resource safety, contract and path completeness and protocol, tests, and user-facing surfaces.