/experiment
Use to run an autoresearch-style experiment loop — improve a measurable artifact by trying changes, measuring against one objective metric, keeping if better and reverting if not, on an isolated git branch with a logged trail. Trigger on "/experiment", "run an experiment loop",
$ npx -y skills add alirezarezvani/gaios --skill experiment --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
- Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.
- You can call itInvoke it directly when you want it.
- Slash command
/experiment
Context preview
The summary Claude sees to decide when to auto-load this skill.
Use to run an autoresearch-style experiment loop — improve a measurable artifact by trying changes, measuring against one objective metric, keeping if better and reverting if not, on an isolated git branch with a logged trail. Trigger on "/experiment", "run an experiment loop",
SKILL.md
experiment.SKILL.mdname: experiment
description: Use to run an autoresearch-style experiment loop — improve a measurable artifact by trying changes, measuring against one objective metric, keeping if better and reverting if not, on an isolated git branch with a logged trail. Trigger on "/experiment", "run an experiment loop", "autoresearch X", "tune/optimize X against a metric", "hill-climb X". Follows references/sops/experiment-loop.md. Enforces the autonomy gate (objective metric + reversible sandbox + no PHI/external) before any unsupervised looping.
Experiment loop (autoresearch)
Hill-climb a measurable artifact: **try → measure → keep-or-revert → log**, on a branch, with a trail. Full procedure: `references/sops/experiment-loop.md`. Existing examples + use-case cards: `experiments/`.
First, the autonomy gate
Before looping unsupervised, confirm **all three**: 1. **Objective, cheap, automatable metric** (code computes it in seconds–minutes). 2. **Reversible sandbox** (git branch / throwaway data / `.tmp`). 3. **No PHI, no confidential data, no external send, no prod deploy.**
If any fails → **human-in-the-loop**: propose each change, get approval, no overnight run. Regulated/production outcomes feed the human-reviewed change process — never auto-deploy. This is `CLAUDE.md`'s scale-caution-to-stakes rule.
Run it
1. Pick or write the experiment's `CARD.md` (objective, metric+direction, the one mutable surface, budget, autonomy, owner). New ones can copy `experiments/forecast-tuning/`. 2. `git checkout -b experiments/<tag>`. 3. Baseline run → `tools/experiment_log.py add <results.tsv> --id ... --metric ... --status keep --note baseline`. 4. Loop: edit the **one** mutable surface → run the **read-only** harness → log → keep (commit) if improved, else `git reset`. Simpler-and-equal = keep. 5. Stop on interrupt or the budget. Hand the winner + trail to the human.
Guardrails
- The eval harness is read-only ground truth — never edit it to game the metric.
- One mutable surface only (reviewable diffs).
- Run artifacts (`results.tsv`, `run.log`, `data/`) are git-ignored; the harness + surface + CARD are the tracked blueprint.
- Don't ship regulated/production changes from the loop.
Read more
name: experiment description: Use to run an autoresearch-style experiment loop — improve a measurable artifact by trying changes, measuring against one objective metric, keeping if better and reverting if not, on an isolated git branch with a logged trail. Trigger on "/experiment", "run an experiment loop", "autoresearch X", "tune/optimize X against a metric", "hill-climb X". Follows references/sops/experiment-loop.md. Enforces the autonomy gate (objective metric + reversible sandbox + no PHI/external) before any unsupervised looping.
Experiment loop (autoresearch)
Hill-climb a measurable artifact: **try → measure → keep-or-revert → log**, on a branch, with a trail. Full procedure: `references/sops/experiment-loop.md`. Existing examples + use-case cards: `experiments/`.
First, the autonomy gate
Before looping unsupervised, confirm **all three**: 1. **Objective, cheap, automatable metric** (code computes it in seconds–minutes). 2. **Reversible sandbox** (git branch / throwaway data / `.tmp`). 3. **No PHI, no confidential data, no external send, no prod deploy.**
If any fails → **human-in-the-loop**: propose each change, get approval, no overnight run. Regulated/production outcomes feed the human-reviewed change process — never auto-deploy. This is `CLAUDE.md`'s scale-caution-to-stakes rule.
Run it
1. Pick or write the experiment's `CARD.md` (objective, metric+direction, the one mutable surface, budget, autonomy, owner). New ones can copy `experiments/forecast-tuning/`. 2. `git checkout -b experiments/<tag>`. 3. Baseline run → `tools/experiment_log.py add <results.tsv> --id ... --metric ... --status keep --note baseline`. 4. Loop: edit the **one** mutable surface → run the **read-only** harness → log → keep (commit) if improved, else `git reset`. Simpler-and-equal = keep. 5. Stop on interrupt or the budget. Hand the winner + trail to the human.
Guardrails
- The eval harness is read-only ground truth — never edit it to game the metric.
- One mutable surface only (reviewable diffs).
- Run artifacts (`results.tsv`, `run.log`, `data/`) are git-ignored; the harness + surface + CARD are the tracked blueprint.
- Don't ship regulated/production changes from the loop.
Clone it, run /setup, and it becomes yours in Claude Code or Codex — a second brain + Chief of Staff that holds your context, structures your work, drafts in your voice, and runs reliable workflows.
Other skills on gaios.
- /audit
Use when someone asks for an AIOS audit, asks to score their setup against the Four Cs, or says "is my AIOS working" / "audit my setup" / "find gaps in my AIOS". Produces a Four-Cs scoreboard with top-3 fixes ranked by leverage.
Open skill - /daily
Use at the start of a working day (or when the user asks "what should I focus on today / give me my brief / daily standup"). Produces a one-screen daily brief — today's top 3, open loops that need a nudge, what's slipping, and a suggested focus order. Trigger on "/daily", "daily
Open skill - /decide
Use when the user faces a real decision with stakes and wants it framed cleanly — options, ranked criteria, a recommendation, the falsifier, and reversibility — then logged to decisions/log.md. Trigger on "/decide", "help me decide", "should I X or Y", "frame this decision",
Open skill - /exec-cockpit
Template skill for a leadership-transition / executive cockpit — when someone steps into or covers a leadership role and needs to not drop anything. Produces a handoff doc + decision-rights map, an "open loops" tracker, a team-comms cadence, and a recurring report/update
Open skill - /graph-ingest
Use when the user brings an external source worth keeping — a URL, a paper, a tweet/thread, a blog post, a docs page — and wants it pulled into the second brain. Fetches the source into the git-ignored capture inbox, admits it to the committed wiki through the admission policy,
Open skill - /graph-query
Use when the user asks a question about how things in the codebase or wiki relate — what connects two things, how something works end to end, what depends on what, or what a concept means in this repo. Answers from the knowledge graph in graphify-out/graph.json, with
Open skill

