| name | menu |
| description | Show what FounderOS can do right now. Say "show me what you can do", "what can FounderOS do", "what should I try next", or "what's relevant right now" (or run /founder-os:menu). Returns 5 to 7 capability suggestions tailored to current state. Reads `brain/.snapshot.md`, open flags, this week's must-do, the last 7 days of `brain/log.md`, and the presence of `core/voice-profile.yml` and `core/brand-profile.yml`. Free-tier accessible - no LLM call inside the algorithm.
|
| why | Shows what the OS can do right now based on your actual state - a static skill list tells you what exists; menu tells you what is relevant today. |
| enhance | Fill cadence/weekly-commitments.md and keep brain/flags.md current - menu scores suggestions against your must-do list and open flags, so empty cadence files produce a generic starter list. |
| allowed-tools | ["Bash","Read"] |
| mcp_requirements | [] |
Menu
Runs on: local-exec - runs a local script; on a cloud surface I read the results, I do not run it.
Discovery entry for FounderOS. Returns a small, ranked list of capabilities the founder is most likely to want right now, scored against current state. Natural-language phrasing is primary, slash commands appear parenthetically only for capabilities that have a real command file.
How this skill runs
The menu engine is scripts/menu.py. The skill invokes the script and prints its output verbatim. The model does not score capabilities, does not invent rows, and does not paraphrase the engine's output.
python scripts/menu.py
The script:
- Reads state files only (
brain/.snapshot.md, brain/flags.md, cadence/weekly-commitments.md, brain/log.md, core/voice-profile.yml, core/brand-profile.yml, context/priorities.md, drafts/).
- Scores capabilities against deterministic rules.
- Returns the top 5 to 7 by score, with a Day-1 starter set as zero-state fallback.
- Renders rows natural-language-first, with slash commands only for capabilities that have a real
.claude/commands/<name>.md file. Skill-only capabilities (weekly-review, priority-triage, pre-send-check) render natural-language only and never invent a slash form.
- Closes with the verbatim line specified in
scripts/menu.py.
- Stdlib only. No LLM call. No network call. Free-tier accessible.
What the skill does
- Locate the FounderOS install root (the working directory, or the repo root if the user is inside a sub-folder).
- Run
python scripts/menu.py --root <root> and capture stdout.
- Print stdout verbatim. Do not summarize, do not paraphrase, do not add commentary.
- If the script exits non-zero, surface the error to the user and stop.
What the skill does not do
- Does not score capabilities itself. The script is authoritative.
- Does not invent slash commands. If the script omits a slash form, the model omits it too.
- Does not write to any file.
- Does not call an LLM. The scoring is deterministic, file-based, free-tier accessible.
Constraints
- Output must end with the closing line from
scripts/menu.py verbatim.
- Trigger phrases the operator already uses ("set up my voice profile", "set up my brand profile", "what should I focus on next", "what's on for today?", "audit the OS") appear in the script's rendered rows when the corresponding capability surfaces. The script handles this; the skill does not.
- If
scripts/menu.py is missing on this install, surface that to the user and stop. Do not fall back to a model-side menu (that would re-introduce the v1.20.0 defect this skill was rewritten to fix).
