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/rsafargalin/sextant/agents-mdgit clone --depth 1 https://github.com/RSafargalin/sextantWrote 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/rsafargalin/sextant/agents-md)<a href="https://agentmods.dev/instructions/rsafargalin/sextant/agents-md"><img src="https://agentmods.dev/badge/instructions/rsafargalin/sextant/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 | $0.01514 | $0.01514 |
| Opus 5 | $0.00757 | $0.00757 |
| Sonnet 5 | $0.00303 | $0.00303 |
| Haiku 4.5 | $0.00151 | $0.00151 |
Grade A, and why
sextant 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 — 122 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — working on sextant
Operational notes for anyone (human or agent) editing this repository. What the tool is belongs in README.md; the iteration history and the reasoning behind decisions belong in docs/. This file is only what you need in order to change the code without breaking something invisible.
Build and verify
swift build && swift test # 330 tests
make ci # build + test + self-lint — run this before committing
swift test --filter "<suite name>" # one suite
make lint runs sextant on its own sources with sextant-rules.json. The tool is dogfooded:
if a change makes lint noisy on this repository, that is a finding about the change.
How a change lands
Every change goes through a branch and a pull request — maintainer changes included. main is
protected: it takes no direct pushes, and a pull request merges only once CI is green on it.
git switch -c fix/what-it-fixes
# … work, `make ci` …
git push -u origin HEAD && gh pr create
The rule is written here because it was broken here. Five commits were pushed straight to main
in one session, the last of them a change across twenty-one files, on the strength of a "commit
straight to main" that had been said once about a one-line release chore. Nothing broke, and that
is not the point: a green local make ci is a predictor, while CI on a pull request is a gate, and
a twenty-one-file change with no reviewable unit is not reviewable in principle.
The protection is what makes this a rule rather than a resolution. It applies to administrators too — otherwise it would guard against everyone except the account that actually pushes here. An emergency turns it off in the repository settings, deliberately and visibly, which is the whole difference from bypassing it by habit.
Layout
Sources/SextantCore— all logic, testable. Anything worth a test lives here.Sources/sextant— thin executable: argument dispatch, rendering, process exit. Keep it thin; the reason the codebase was refactored once already is that 1257 lines ofmain.swiftwere untestable.Tests/SextantCoreTests— behavioural tests, plus a fixture package underTests/Fixtures/IndexFixtureused to exercise a real IndexStoreDB.plugins/sextant— the Claude Code plugin: the MCP registration, a skill and a command. It ships no code of its own beyond two launcher scripts, registers no hooks (a hook that writes from the moment it is installed is the user's decision, not an install's), and it must not restate the MCP tool list —MCPTools.definitions()is that list, and the skill describes when to ask rather than what exists..claude-plugin/marketplace.jsonat the root is what serves it.
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 · 122 lines · 1,514 tokens per session scan A ba0f09ab334a
sextant AGENTS.md is an instructions file published in the GitHub repository RSafargalin/sextant (2 stars, last pushed 16d ago), licensed Apache-2.0. It adds 1,514 tokens to every session, about $0.0076 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
repobrain copilot-instructions.md
Copilot instructions for study8677/repobrain, covering github copilot bootstrap instructions and hard rule — query the repobrain hub first.
Swift-Concurrency-Agent-Skill copilot-instructions.md
Instructions for AvdLee/Swift-Concurrency-Agent-Skill, covering github copilot instructions: agent skills expert, core agent skills principles, format compliance, skill content organization and review focus areas.
Swarm AGENTS.md
Instructions for christopherkarani/Swarm, covering agent instructions, public repository hygiene and development rules.
liney AGENTS.md
Instructions for everettjf/liney, covering liney repository collaboration guide, 交流语言 / communication language, project overview, repository layout and build and test.
skills-manager CLAUDE.md
Instructions for yibie/skills-manager, covering skills manager - development guide, tui engine: blessed, 核心原则, 参考文档 and 正确的 list 用法.
swift-secp256k1 AGENTS.md
Instructions for 21-DOT-DEV/swift-secp256k1, covering agents.md (swift-secp256k1), commands, non-obvious patterns, code style and boundaries.