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/as-foss/mandible/agents-mdgit clone --depth 1 https://github.com/AS-FOSS/mandibleWrote 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/as-foss/mandible/agents-md)<a href="https://agentmods.dev/instructions/as-foss/mandible/agents-md"><img src="https://agentmods.dev/badge/instructions/as-foss/mandible/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 | $0.04741 | $0.04741 |
| Opus 5 | $0.02371 | $0.02371 |
| Sonnet 5 | $0.00948 | $0.00948 |
| Haiku 4.5 | $0.00474 | $0.00474 |
Grade A, and why
mandible 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 today.
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 — 361 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — working agreements for AI agents in this repo
Read this before touching code. Every rule a script can check is a pointer to that script. What is left is prose, because no lint can catch it.
Precedence: docs/design.md is the design authority, so it says what to build and
why. This file is the operational authority, so it says how to work here.
CONTRIBUTING.md is for humans. If this file and docs/design.md disagree about
design, docs/design.md wins and you should fix this file.
1. The invariant that defines the project
No per-tool logic, ever. No
if tool == "docker", no tool-name-keyed special case in any extraction tier, no per-tool patch file vendored into this repo.
Tool-specific knowledge lives in exactly one place: user-local override files
under ~/.config/mandible/overrides/, which are never committed here. (Spec
revision 3 deleted the vendored catalog that used to be the second place — a
per-tool catalog is per-tool knowledge relocated into data, and it cannot stay
current with the tool actually installed. Parsing is keyed by framework now:
see spec §7 Tier A′ and mandible-extract/src/help_text/profile.rs, where
adding a framework is one match arm plus one fingerprint.)
If a tool renders badly, the fix is a better general parser, a new general tier, or an honest low-confidence badge in the UI. It is never a special case. This is the entire reason the tiered architecture exists; one exception starts the erosion that the architecture was built to prevent.
2. Hard limits
Every number here is enforced, and none of them is a law of nature. Changing one is a decision, and a decision goes in docs/design.md §16.
| Limit | Number | Enforced by |
|---|---|---|
Code lines in a .rs file before mod tests |
800 | scripts/shape_guard.sh |
| Lines in a function | 150 | clippy.toml, clippy::too_many_lines |
| Cognitive complexity of a function | 30 | clippy.toml, clippy::cognitive_complexity |
| Lines in one comment block | 12 | scripts/shape_guard.sh |
| Comment lines over code lines in a file | 0.5 | scripts/shape_guard.sh |
| Tools a new parser heuristic must move | 5 | xtask sweep-diff, read by a human |
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.
- today Changed · -100 lines · -3,186 tokens per session scan B → A e96e1d2a64ae
- 5d ago First seen · 461 lines · 7,927 tokens per session scan B 6eab7ac79a3d
mandible AGENTS.md is an instructions file published in the GitHub repository AS-FOSS/mandible (141 stars, last pushed today), licensed Apache-2.0. It adds 4,741 tokens to every session, about $0.0237 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
anarlog AGENTS.md
AGENTS.md instructions for fastrepl/anarlog, covering commands, pre-commit verification, guidelines, code style and cli tui command architecture.
warp AGENTS.md
AGENTS.md instructions for warpdotdev/warp, covering agents.md, development commands, build and run, running with local warp-server and connect to server on default port 8080.
psmux AGENTS.md
AGENTS.md instructions for psmux/psmux, covering agent instructions and test isolation.
terminal-browser AGENTS.md
AGENTS.md instructions for zenbu-labs/terminal-browser, a project described as: A browser inside your terminal.
VTCode AGENTS.md
AGENTS.md instructions for vinhnx/VTCode, covering agents.md, rules, detailed guides, workspace and per-module guidance.
pisper AGENTS.md
AGENTS.md instructions for ling-kong-ran/pisper, covering agents.md, project, architecture, commands and setup.