External LLM invocation. Triggered ONLY by @council,@probe,@crossref,@gpt,@gemini,@grok,@qwen.

Local Development Standards define how developers set up and run projects on their machines, ensuring consistency, reproducibility, and minimal onboarding friction.

Declare variables in smallest possible scope, initialize close to first use, minimize span and live time

Archives completed logs to cloud storage with index management and cleanup

Orchestrates single-log sequential workflows by coordinating operation skills for individual logs

log-observability-issues

Sync Logseq pages tagged with #zotero to Zotero by adding 'in_logseq' tag. Uses batch checking for efficiency. Claude Code only.

Lomb-Scargle periodogram for finding periodic signals in unevenly sampled time series data. Use when analyzing light curves, radial velocity data, or any astronomical time series to detect periodic variations. Works for stellar rotation, pulsation, eclipsing binaries, and general periodic phenomena. Based on lightkurve library.

Manage long-running agent sessions. Use for tracking progress in extended tasks, maintaining context across long sessions, and managing multi-step workflows.

Manage Loom team tickets as a team manager with limited permissions (no git merge/push, no loom team commands). Handle ticket lifecycle: create, update, add notes, track dependencies, identify blockages.

Act as Team Manager for Loom team, handling ticket workflow when team commands may be unavailable

>

>-

Compare simulation results between Driftfusion and other simulators (ASA, IonMonger), analyze discrepancies in J-V characteristics, and configure simulations for fair comparison.

Close the loop on a completed plan by writing SUMMARY.md and updating state

Citation verifier — checks that in-text citations accurately represent what the cited source actually says

first step of planning

meme-launcher

Weekly experiment tracking loop for MD Home Care. Scans content changes, measures traffic impact via PostHog and GSC, and makes keep/iterate/revert decisions with lag-adjusted attribution.

>

>

Audit codebase files against the 4-pillar quality manifesto using RECON work batches, with batch processing and context budget management (Titan Paradigm Phase 2)

Import, validate, query, and manage job aids. Job aids are structured condition/action lookup tables that parameterize how a task executes based on conditions like loan program, income type, and state. Use when working with decision tables, lookup matrices, or conditional business rules.

>-

>

Deep research on a technical topic for blog writing. Use when user says research, look into, explore, or provides a blog topic to investigate. Searches the web extensively and creates structured research notes with visual opportunity identification.

pdca

Follow the instructions in ./workflow.md.

Run four phases in order. Each phase is conversational and inline — no

Track bugs, QA, and fixes in .galdr/. For bug reports, issues, quality, fixes, QA workflows.

Track machine learning experiments with reproducible parameters and metrics

Assess dependency health across CVE exposure, freshness, maintainer risk, and license compatibility

Follow the instructions in ./workflow.md.

Deploy code from the current directory to Railway using `railway up`.

Creates or removes a git worktree using the project scripts and shared config. Use when the user asks to create a worktree, add a worktree, remove a worktree, delete a worktree, or work on a parallel branch in a separate directory.

Privacy-protected ID management with @ledgerhq/client-ids — DeviceId, UserId, DatadogId, export-rules.json

>-

Draft cold emails, warm intro blurbs, follow-ups, update emails, and investor communications for fundraising. Use when the user wants outreach to angels, VCs, strategic investors, or accelerators and needs concise, personalized, investor-facing messaging.

>-

Docker containerization best practices. Use when creating Dockerfiles or docker-compose configurations.

Assign task risk tier and change type with explicit rationale.

>-

>-

When explicitly authorized, execute end-to-end with a phased plan in one run until Definition of Done is met.

Manage AI agent skills across projects — sync, validate, and maintain shared skill libraries.

>-

Validar a politica de remotes do repositorio antes de qualquer commit/push; operar em single remote por padrao e usar dual remote apenas quando o PO determinar.

Follow the instructions in ./workflow.md.

Write or refresh `PLANS.md` for the current task in a way that matches repository policy.