Data & Analytics

Fix silent metadata issues in RFDB node storage. Covers two traps: (1) metadata flattening — nested metadata fields become top-level after serialization, so node.metadata.field is undefined but node.field works. (2) reserved keys — fields named "type", "id", "name", "file", "exported" are silently stripped from metadata by _parseNode() to prevent overwriting top-level node fields.

Fix stale Haskell analyzer binaries silently being used by the orchestrator even after a successful rebuild. Use when: (1) modified `packages/js-analyzer/` or other Haskell analyzer source, ran `scripts/build-native.sh ... cabal install`, but a subsequent `grafema analyze` shows old behavior (missing new edges, no new node types, etc.), (2) `~/.cabal/bin/grafema-analyzer` has fresh timestamp but `grafema analyze` produces output as if the change didn't happen, (3) a test verifying new analyze...

Extract knowledge (decisions, facts, session metadata) from the current Claude Code session into the Grafema Knowledge Base. Run after completing a task or at any point when substantive knowledge was produced. Follows runbook _ai/runbooks/02-claude-sessions.md.

Fix intermittent `{:no_translation, :unicode, :latin1}` crashes in Elixir escript daemons that use length-prefixed framed IPC on stdin/stdout. Use when: (1) daemon worker crashes only on some input files, usually ones with non-ASCII bytes (kanji, cyrillic, emoji); (2) error surfaces as `Protocol error` from daemon's error branch or as garbled frame-length bytes seen by the orchestrator/client side; (3) standalone one-shot mode works fine on the same input but multi-request daemon mode fails; ...

Fix Node.js CLI tools crashing inside Docker containers when host-installed node_modules require a newer Node version than the container provides. Use when: (1) "SyntaxError: Invalid regular expression flags" with /v flag in string-width or similar packages, (2) node_modules installed on host with Node 20+ but container has Node 18, (3) `npm install` with file: protocol creates symlinks that break inside Docker, (4) pnpm workspace packages become broken symlinks in containers. Covers version ...

Fix docker exec hanging when starting background processes (servers, daemons) inside containers. Use when: (1) docker exec never returns despite using & or nohup, (2) background server started in container causes docker exec to hang indefinitely, (3) env_startup_command in SWE-bench or similar frameworks times out, (4) setsid/disown needed for proper process detachment in Docker. Root cause: docker exec tracks ALL processes in the exec session, not just the top-level PID.

Fires after fixing any non-trivial bug or regression. Asks: could the graph have caught this as a guarantee? Pairs with reflection-in-and-on-action — picks up after "earliest catchable signal" and asks the next question: was that signal expressible in graph? Triggers: (1) after any non-trivial bug fix is verified working, (2) after a regression report (something used to work, broke), (3) during step 6 (knowledge extraction) of the workflow, (4) when reflection-on-action surfaces a "would have...

Fix Elixir/Erlang AST processing bugs in Grafema beam-analyzer. Use when: (1) Elixir parser returns MODULE node but 0 functions/calls — body nesting issue, (2) Erlang parser crashes with "cannot convert list to string" on OTP 26+ — location format changed from integer to keyword list, (3) pipe operator |> creates spurious CALL nodes instead of desugared function calls — clause ordering bug, (4) multi-module .ex files return only the first module — missing __block__ handler, (5) installing Erl...

User says `/approve` after reviewing a completed task in a worker worktree. This replaces the manual merge-to-main workflow.

Reads a CARLA Leaderboard results.json — per-route driving score, route completion, infraction penalty and the infraction lists — explains how the score was computed for that leaderboard version, and rescores an existing result under the other version's rules (2.0 multiplicative vs 2.1 additive) without re-running anything. Also diffs two runs and merges sharded result files. Use when the user asks "what did I score", "why is my score low", "did my agent pass", "why don't my scores match the ...

Structure information so people can find what they need, understand where they are, and navigate confidently. Covers navigation pattern design, taxonomy, labeling systems, search and browse strategy, wayfinding, and IA research methods. Trigger when designing navigation structures, categorization schemes, site maps, taxonomies, labeling systems, search experiences, or asking "how should we organize this?" Also trigger for card sorting, tree testing, information findability problems, or when u...

Charts, evidence tables, process/argument diagrams, and figure validation for research reports and direction discussions. Load when producing or reviewing any long report, when a direction discussion needs implementation-chain/risk/decision diagrams, or when the user asks to plot data, draw diagrams, or produce publication-ready figures. Delegates statistical analysis and verdicts to result_analysis.

Evidence interpretation, baseline-expected-actual comparison, mechanism diagnosis, and statistical verdicts. Load when the user asks to analyze results, interpret data, decide whether to kill/pivot/continue a direction, or evaluate numeric evidence. Locates which causal link failed, enforces convergence accounting, and delegates chart generation to result_visualization.

Research idea convergence advisor. Load when the user proposes a new research direction, asks "is this idea worth pursuing", requests literature search / gap analysis, or needs stop criteria for a study. Grounds ideas in reality — implementation closure, related-work facts, speculation tracking, decision-driven experiments — before any formal experiment design. Experiment execution itself belongs to experiment_manager.

Publish the current R Weekly draft as a new weekly post. Use when an editor invokes /release or asks to "release", "publish", or "create the weekly post" from draft.md. Validates the draft before creating the file in _posts/.

Generate the R Weekly Highlights Poll Slack text. Parses draft.md for all links, presents them for editor selection, then formats two /poll Slack commands (5 items each). Use when an editor wants to create the highlights poll for the #highlights Slack channel.

Find, download, resize, and embed images for the three Highlight links in draft.md. Fetches each article, picks a representative image, runs upload_image() to push to the rweekly/image repo, then inserts the image embed below each highlight link in draft.md. Has human-in-the-loop review gates before downloading, before pushing to GitHub, and before editing draft.md.

Guide the R Weekly curation team through preparing a new weekly issue. Reads curatinator_latest.md, classifies posts by section, checks for duplicates, and populates draft.md. Use when an editor wants to process the weekly content pipeline.

ML and data science workflows — EDA, feature engineering, modelling, evaluation, and production handoff. Use when exploring data or building models.

Handle Google Sheets comment review, replies, resolves, reopens, and notification emails correctly. Use this whenever the user asks to process, review, reply to, resolve, reopen, audit, verify, or summarize Google Sheets comments; mentions Google Sheets comment notification emails; asks to handle spreadsheet feedback; or opens/inspects a Google Sheet specifically to deal with comments. Do not use for ordinary spreadsheet editing or non-comment data analysis. This skill prevents treating Gmail...

Handle Google Sheets comment review, replies, resolves, reopens, and notification emails correctly. Use this whenever the user asks to process, review, reply to, resolve, reopen, audit, verify, or summarize Google Sheets comments; mentions Google Sheets comment notification emails; asks to handle spreadsheet feedback; or opens/inspects a Google Sheet specifically to deal with comments. Do not use for ordinary spreadsheet editing or non-comment data analysis. This skill prevents treating Gmail...

Vet, score, or decide whether to add Twitter/X accounts to a curated list. Use when asked to vet a handle, score accounts, decide whether an account belongs on a list, or when twitter-digest/twitter-discover need shared account scoring.

Vet, score, or decide whether to add Twitter/X accounts to a curated list. Use when asked to vet a handle, score accounts, decide whether an account belongs on a list, or when twitter-digest/twitter-discover need shared account scoring.

Guide for working with Feast (Feature Store) — defining features, configuring feature_store.yaml, retrieving features online/offline, using the CLI, and building RAG retrieval pipelines. Use when the user asks about creating entities, feature views, on-demand feature views, stream feature views, feature services, data sources, feature_store.yaml configuration, feast apply/materialize commands, online or historical feature retrieval, or vector-based document retrieval with Feast.