autoresearch: a skill-only autonomous experiment loop — propose → measure → keep winners → repeat, against any numeric objective, with noise-floor gating so nothing is banked on noise. /ar:start, /ar:resume, /ar:status, /ar:report, /ar:stop. State stays in the target repo's ./.ar/. No hooks, no MCP.
Plugin ar from marketplace ccplugins in emaballarin/ccplugins. Commands read from the repo's manifest — 17 days ago.
reportWrite the final summary of an autoresearch run to `./.ar/final_report.md` — trajectory, best commit and its diff, what worked, what failed, and what to try next. Use for `/ar:report`, "write up the autoresearch results", or "summarise the experiment loop". Reads `./.ar/ar.jsonl` from disk, so it wor
resumeRun the autoresearch experiment loop — propose one change, measure it, keep it only if it beats the noise floor, repeat. Use to resume or continue an autoresearch run, for `/ar:resume`, for "keep iterating", "next experiment", "continue optimising", or whenever `./.ar/ar.jsonl` exists and the loop s
startOpen a new autoresearch experiment loop — dedicated `ar/…` git branch, `./.ar/` state, and a measured baseline noise floor so later gains can be told apart from run-to-run scatter. Invoke ONLY when explicitly asked to start an autoresearch run, open an experiment loop, or `/ar:start` — e.g. "autores
statusPrint the current autoresearch run's state — goal, baseline, noise floor, best-so-far, run tallies, budget, branch, and the last few iterations. Use for `/ar:status`, "how is the loop doing", "what's the best so far", or any read-only check on an autoresearch experiment. Reconstructs everything from
stopEnd the current autoresearch run — append a `status:stopped` sentinel to `./.ar/ar.jsonl`, which is also what halts the external `ar-loop.sh` driver. Use for `/ar:stop`, "stop the loop", "halt autoresearch", or "that's enough iterations". Writes the sentinel and the final report, then prints (does n
Research & scientific-computing skills adapted from Claude Science for stock Claude Code: literature-review, bib-audit, pdf-explore, figure-style, figure-composer, paper-narrative, paper-review, plus computational-scientist and deep-researcher subagents.
7 skills · 2 agents
mindfunnel: project-agnostic Claude Code session management. /mf:dump consolidates session state into auto-memory; /mf:spinup reads it back; /mf:park and /mf:unpark carry the session scratchpads across a reboot; /mf:setup and /mf:prime handle first-time and per-project scaffolding; /mf:author is the reference for writing the documents an agent reads — skills, AGENTS.md, CLAUDE.md.
7 skills
tuneml: a read-first advisor for training and evaluating neural networks well — the scientific method for tuning and the speed↔quality frontier in one place. /tml:audit ranks graded, priced findings across the algorithmic, systems, protocol and architecture tiers; /tml:plan fixes an operating point and a step budget; /tml:round designs the next experiment, splitting hyperparameters into scientific, nuisance and fixed; /tml:analyze runs the diagnostic checklist and returns a variance-aware adopt verdict; /tml:review adversarially checks a change. Runs local or remote, and analyses results it did not produce. Supersedes parml. No hooks, no MCP.
5 skills
whetstone: sharpen the thinking before the work, and the work before it lands. /ws:grill (user-only) interviews a plan or decision in rounds over its design tree until nothing is silently assumed; /ws:test-gate (model-invoked) holds every new test to a shared value bar and extends rather than restructures existing ones; /ws:test-audit (user-only) audits and consolidates an existing suite, evidence first; /ws:deslop (user-only) strips AI slop from a diff before /code-review; /ws:review-gate (model-invoked) puts a change and its intent — not its reasoning — in front of a fresh-context reviewer before it is committed, released or run. No hooks, no MCP.
5 skills
dm42: write, review and debug HP-42S-family RPN programs, the SwissMicros DM42n first (also DM42, Free42, Plus42). /dm42:rpn is the workflow and the verified reference — program-text rules, stack lift, flags, locals, menus, GETKEY vs GETKEY1, PIXEL/AGRAPH, the DM42 virtual variables and graphics modes — with a round-trip check for the lines the text parser drops silently. /dm42:setup builds dm42sim: SwissMicros' own Free42 core compiled as for the device, driven headless by action scripts that return JSON. No hooks, no MCP.
2 skills
ccbar: a quiet status band above the prompt, drawn natively by Claude Code's function hooks — model and effort, repository (GitHub owner/name and branch, optionally as links; diff counter), a context bar coloured by the /context categories with used/total and load, the session's new tokens (cache reads of the re-sent conversation left out, or counted too when set) counted once per API response, and the session and weekly rate limits with reset countdowns. Colour carries meaning; refreshes are paced. Function hooks only: no command hooks, no MCP, no settings patching.
Talk like caveman. Cut filler. Keep technical details explicit.
22 skills · 1 command · 3 agents · hooks
Persistent memory for Claude Code: the next session starts where the last one left off.
29 skills · 1 MCP server · hooks
tasteskill 2.5.1: professional company websites at the figma.com and x.ai level, with one section per screen, real photographs, real product UI, effects kept for tech and gaming pages, and a real-browser checker and finish gate.
1 skill
Open-source AI job search that runs locally in your AI coding CLI: scan job portals, evaluate listings into a structured A-H report with a 1-5 score, tailor your CV, track applications
5 skills · hooks
Cross-session memory and token savings for coding agents.
6 skills · 1 agent · 1 MCP server · hooks
Package any website or local web build into a lightweight desktop app with Pake CLI.
1 skill
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.