最近新增
刚加入收藏的新技能，抢先试用。
plain-prose
polycode-public
Holds this repo's writing rules for plain, human prose and the LLM-voice tells to cut. Load it before writing any human-facing text — docs, code comments, reports, runbooks, site copy, or chat replies.
iterate
polycode-public
Run the delivery cycle unattended until the board has no machine-only row that can start — board, a wave of isolated sub-agent batches on one branch and one PR, watch, auto-merge, watch, board, again. A Plan–Do–Check–Act cycle in the orchestrator–workers shape. Invoke when the operator says "iterate", "run the cycle", or "keep going until the board is clear".
do-next
polycode-public
Work NEXT.md top to bottom as waves of concurrent worktree sub-agents, land them on one batch branch, push in batches, raise one PR, and hand over to /watch. Invoke when the operator says "do next", "work the board", "clear NEXT.md", or when a landed batch leaves rows still open.
cool-down
polycode-public
Slow the flow of new work and settle what is already in flight, then bring it back up again. Agents commit and stop, PRs and workflow runs are driven green one branch at a time, and the tracking documents catch up. Holds the revival notes too, so /wake reads from here. Invoke when the operator says "cool down", or when a batch is stacking problems faster than it lands them.
compact-ready
polycode-public
Before a compaction, land or commit every agent's and worktree's work, write NEXT.md and chat decisions to files, keep monitors, then answer yes / yes after / not yet and list what the summary must keep. Invoke as /compact-ready; /board and /auto-merge run it at their natural breaks.
clean
polycode-public
Find what this repository has left behind (failed stacks in the sandbox, merged branches, worktrees of merged branches, logs and test artefacts, build output) and remove every category the operator agrees to in one go; then, when nothing is in progress, fetch, switch to main and pull. Invoke when the operator says "clean", "clean up", or "tidy the repo".
capabilities
polycode-public
Find the tooling this repository already has before building any, and keep REPORT_CAPABILITIES.md current. Invoke before you add a script, workflow, stack, check, test harness, ingest step, eval, page or skill; when a task says add, build, create, automate, check, ingest, measure, deploy or clean up; and when you need the command that does something here.
board
polycode-public
Render the open-work board — one table for every NEXT.md row in board order, then every PLAN_*.md in one of three sub-parts (open tasks not yet on the board, every open task boarded, nothing open: an archive candidate), the deployed stacks in polycode-sandbox, the workflow runs on main and every open PR, and a branch audit — then write the sequenced board back to NEXT.md. Ends with /compact-ready. Invoke when the operator asks for the board, the open items, or "what's in flight".
auto-merge-dry-run
polycode-public
Everything /auto-merge reports, with nothing it changes — the same catalogue, tables and verdicts, and no merge, commit, push or file write. Invoke when the operator wants to see what would merge before anything does.
auto-merge
polycode-public
Catalogue worktrees, branches, uncommitted work, PRs and review threads; render the state; merge every PR that is ready; then hand over to /watch. With the argument "dry-run" it changes nothing and prints what it would do. Invoke when the operator says "auto-merge", "merge what's ready", or asks for the merge state.
hello-review
andrewjr1991
Read-only smoke test that summarizes the current project and suggests one safe improvement.
flyrank-data
omar320250068-prog
The FlyRank internship datasets — the 30k-row starter CSV and its gotchas, the ~79M-row warehouse release tables and grains, panel warnings, access, and iteration rules. Load for EVERY task that touches the data. (Project-specific: delete this folder when reusing the skill library elsewhere.)
flyrank-context
omar320250068-prog
FlyRank-specific background — what the company does, the search problem, and the hand-written flags story that the ML internship work builds on. Load for context when an assignment references FlyRank's product, flags, or the "beat the rule" framing. (Project-specific: delete this folder when reusing the skill library elsewhere.)
writing-research-papers
omar320250068-prog
Structures and writes a public research page — canonical sections (abstract through limitations), storytelling that carries findings, chart hygiene, referencing for credibility, repurposing for different audiences. Use when turning finished analysis into a paper, portfolio page, or public write-up.
writing-honest-claims
omar320250068-prog
Writes findings in language the evidence can carry — the claim ladder (observed → directional → decision-support, never causal without a design), effect sizes over drama, banned phrasings. Use when writing any conclusion, report section, playbook, or public page from analysis results.
writing-data-contracts
omar320250068-prog
Writes a data contract — what a row means, which fields are features vs labels vs context vs excluded, over which time windows — and verifies every claim with a query. Use before any feature building or modeling, or when results look wrong and the data definition is suspect.
training-honest-models
omar320250068-prog
Trains a first model the honest way — method chosen to fit the question, compared against the baseline on the same split and metric, errors read before scores are believed. Use when moving from a rule baseline to a learned model, or when reviewing a model that reports only a single score.
querying-big-datasets
omar320250068-prog
Works with datasets far too big to download or load in pandas — SQL over remote Parquet with DuckDB, aggregate-then-model, iterate on samples. Use when a dataset has millions of rows, lives on a remote host (hf:// or s3), or a notebook runs out of memory or hits rate limits.
hunting-leakage-and-validating
omar320250068-prog
Finds label leakage and designs honest validation — leakage taxonomy, grouped and time-aware splits, base rates, and the attack-your-own-model checklist. Use before trusting any metric, when a score looks too good, or when features and labels share time windows or origins.
framing-ml-problems
omar320250068-prog
Frames a data/ML problem before any modeling — the decision, the action, the cost of a wrong call, task type, target, and success metric. Use when picking a project direction, writing a research question, or mapping a problem onto an ML task type.
directing-your-ai-assistant
omar320250068-prog
Teaches the working method for using an AI coding assistant on analysis tasks — one task per fresh conversation, search before assuming, interview before big plans, validate everything, commit per finished task. Use at the start of any assignment, whenever a conversation gets long and confused, or when the assistant starts guessing.
deploying-static-pages
omar320250068-prog
Deploys a static page (research paper, portfolio piece) for free from a GitHub repo using GitHub Pages — setup, file layout, verification, and recording the final URL. Use when publishing a finished write-up as a public web page.
building-baselines
omar320250068-prog
Builds the transparent rule-based baseline every model must beat — a hand-written score with reason codes, ranked output, and precision@K evaluation. Use before training any model, or when someone reports model results with nothing to compare against.
auditing-signals
omar320250068-prog
Runs an honest EDA and signal audit — distributions first, heavy-tail handling, verdict-style mini-tests (CONFIRMED/OPPOSITE/MIXED/FALSE), sample-size floors. Use before modeling, when testing whether a popular belief shows up in data, or when a correlation looks too good.