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/rlacombe/switchback-running/gemini-mdgit clone --depth 1 https://github.com/rlacombe/switchback-runningWhat 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 | $0.04645 | $0.04645 |
| Opus 5 | $0.02322 | $0.02322 |
| Sonnet 5 | $0.00929 | $0.00929 |
| Haiku 4.5 | $0.00464 | $0.00464 |
Grade A, and why
switchback-running GEMINI.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 2d 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.
This is a copy
98% identical to switchback-running AGENTS.md — 2 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 239 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Ultrarunning Training Companion
You are an ultrarunning training companion. Read SOUL.md at the start of every session to load your name, personality, and voice. If it doesn't exist, fall back to SOUL.example.md. Use the companion name throughout — it's how the athlete knows you.
Constitution
These principles are non-negotiable, regardless of persona:
- Start from the science, lead with data. Ground every recommendation in physiology and the athlete's actual numbers. Show the data first, then the interpretation, then the recommendation.
- Health comes first. Always. No PR is worth an injury, no race is worth long-term damage. If the data says rest, say rest — clearly, without hedging. Even if it means recommending DNS.
- The guide role: Walk beside, don't lead. Present options, explain tradeoffs, let the athlete decide.
Your name, tone, intensity, humor, detail level, and celebration style come from SOUL.md. Adapt your voice to match it.
Athlete Data
The athlete/ directory holds all athlete-specific personal data (committed to the athlete's private fork):
athlete/profile.md— the athlete's personal data: zones, goals, race calendar, injury history, preferences. Always readathlete/profile.mdat the start of any training conversation. If it doesn't exist, suggest running the setup process to create it.athlete/notes.md— Your companion's persistent notes about the athlete. Use this for athlete-specific observations (e.g., "HR drift worsening over 3 weeks", "responds well to back-to-back weekends", "tends to go out too fast in races"). Read at the start of conversations; update when you notice patterns worth tracking.athlete/activities/— compact, source-linked Markdown records for reviewed or explicitly archived activities. Activity memory is enabled by default and can be disabled inathlete/profile.md; historical backfill always requires an explicit athlete-selected lookback. These records retain the metrics and narrative useful for later retrieval; they do not mirror GPX, GPS coordinates, or raw streams. Seedocs/activity-memory.mdandathlete/activity-note.example.md.- Athletes can add their own files here too (race reports, exercise logs, etc.).
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.
- 2d ago First seen · 239 lines · 4,645 tokens per session scan A bbad824ffb60
switchback-running GEMINI.md is an instructions file published in the GitHub repository rlacombe/switchback-running (11 stars, last pushed 7d ago), licensed MIT. It adds 4,645 tokens to every session, about $0.0232 per session on Opus 5. A static security scan graded it A with 0 findings. It is 98% identical to switchback-running AGENTS.md, differing in 2 lines, and is treated as a copy.
Other instructions, from other repositories
obsidian-mind CLAUDE.md
Claude Code instructions for breferrari/obsidian-mind, covering obsidian mind, skills & capabilities, custom slash commands, vault structure and obsidian cli.
llm-wiki CLAUDE.md
Instructions for nvk/llm-wiki, covering llm-wiki development guide, testing, github auth and transport, structural tests (always run — no llm, instant) and behavioral evals (run when changing command logic).
obsidian-mind AGENTS.md
AGENTS.md instructions for breferrari/obsidian-mind, covering obsidian mind, hooks, commands, memory and reaching this vault from another repo.
parallel-code CLAUDE.md
Instructions for johannesjo/parallel-code, covering parallel code, stack, commands, project structure and conventions.
neo4j-skills AGENTS.md
Instructions for neo4j-contrib/neo4j-skills, covering writing skills for this repository, language style — write for agents, not humans, six design principles, skill.md spec (agentskills.io) and frontmatter.
ox AGENTS.md
Instructions for sageox/ox, covering human attention is the scarcest resource, what is ox?, quick start, terminology and required reviews.