Stats
Actions
Tags
From autoresearch
Resume or continue an active autoresearch experiment loop. Use when autoresearch.md exists, or to resume after a pause/context reset.
How this skill is triggered — by the user, by Claude, or both
Slash command
/autoresearch:resumeThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
You are resuming an active autoresearch experiment loop after a context reset or pause.
You are resuming an active autoresearch experiment loop after a context reset or pause.
autoresearch.md — understand objective, metrics, files in scope, Problem Profile, Decision Tree, what's been tried, landscape model, tabu list${CLAUDE_PLUGIN_ROOT}/scripts/status.sh — see dashboard with keep rate trends and improvement curveautoresearch.jsonl — last 20 lines. Note the keep rate of the last 20 experiments. If < 15% keeps, you're likely converging — focus on layer changes, not micro-opts.autoresearch.ideas.md exists, read it fully. Focus on untried ideas. Prune stale entries.git log --oneline -n 10 — see recent experiment commitsmdvault search "autoresearch technique <bottleneck-type>" --top-k 5. Also run mdvault search "autoresearch anti-pattern <bottleneck-type>" --top-k 5 to avoid known dead ends. Add relevant findings to autoresearch.ideas.md.Read the Loop Rules from ${CLAUDE_PLUGIN_ROOT}/skills/create/SKILL.md. Each iteration: edit code → prepare-experiment.sh → run-experiment.sh → log-experiment.sh. If keep rate < 15% or you feel stuck, also read ${CLAUDE_PLUGIN_ROOT}/skills/create/REFERENCE.md for exploration strategies and patterns. Start the next experiment immediately.
LOOP FOREVER. NEVER STOP. NEVER ASK "should I continue?"
Be careful not to overfit to the benchmarks and do not cheat.
npx claudepluginhub sderosiaux/claude-plugins --plugin autoresearchCore autonomous research loop. Reads research.md, proposes hypotheses, runs experiments, evaluates results mechanically, keeps improvements, discards failures, and iterates until the target metric is achieved or the iteration budget is exhausted. TRIGGER when: user invokes "autoresearch" (no subcommand); research.md exists; user wants the 5-stage loop; user wants iterative optimization overnight.
Resumes paused code experiments by listing them, checking out git branches, loading configs/results history/git logs, summarizing progress/metrics/patterns, and prompting for next iteration or loop.
Runs iterative experiments to optimize measurable metrics (speed, accuracy, config). Manages .lab/ directory for experiment history and autonomous workflow.