DazzleML
- 14 skills
- 0 followers
- 6 hours ago last updated
- ▌ Oracle · dazzlemlKnowledge oracle — query the project's design docs, MOCs, postmortems, GitHub issues, and code to get traced, sourced answers. The entry point when you need to understand how anything in the project works or connects.
- ▌
- ▌ Wherearewe · dazzlemlInbound context recovery: find existing WhereWeAre docs, postmortems, issues, and git state to answer 'what's going on in this project?' Companion to /whereweare (which writes snapshots).
- ▌ Whereweare · dazzlemlForward-looking project snapshot: what was done, what's next, key files, open issues, related projects, and everything needed to resume work after time away.
- ▌ Familiarize · dazzlemlRebuild project context at session start by reading key docs, git state, issues, and design history. Accepts an optional argument describing the current focus or goal.
- ▌ Investigate · dazzlemlStart a structured investigation for new work — gather context from files, issues, git, and agents, then synthesize into design docs, create a plan, and file GitHub issues. Accepts a goal description with hints about information sources.
- ▌ Double Check · dazzlemlExhaustively validate every claim, number, calculation, code-fact, and command in a document or outbound communication. Extracts all claims into a prioritized ledger (flagging what is NOT covered), validates each by the right method (run the math, run the command, cite the file:line, quote the primary source), then calibrates certainty with CIA Words of Estimative Probability and a provenance tier — and proposes corrected, properly-hedged rewrites for any overclaim. Use before sending a message, filing an issue, committing analysis, or whenever you state exact numbers/assertions others will rely on.
- ▌ Collabn Local · dazzlemlRun an N-round expert consultation with one of our own sub-agents (default: brainstorm) that has repo access -- like collaborate3, but the consultant can read the code, trace design docs, and run real tests. Spawn once, continue via SendMessage each round; saves 2N+1 verbatim round files plus an independent dev-workflow final assessment. N defaults to 3.
- ▌ Repokit Setup · dazzlemlSet up a new project using DazzleTools git-repokit-template and git-repokit-common. Creates repo, adds subtree, initializes standard structure.
- ▌ Create Project · dazzlemlFull GitHub project creation workflow: repo from template, subtree, hooks, versioning, private init, issues, topics, discussions, traffic tracker, PyPI setup, and first release.
- ▌ Test Checklist · dazzlemlProduce a hand-runnable human test checklist alongside automated tests. Run after writing automated test suites for a phase/feature that ships user-visible behavior. Creates a markdown doc (public by default in tests/checklists/, or in a project-private location for sensitive content) that a human walks through to verify end-to-end behavior that mocks can't cover -- shell rendering, Windows codepages, real subprocess behavior, interactive prompts, file atomicity under real conditions, config isolation, and other "the mocks say it works but does it REALLY" scenarios.
- ▌ Merge 3 Way Split · dazzlemlRead-only pre-merge review. Lists dual-touched files, shows OURS / THEIRS / merge-base diffs, and flags likely-collision patterns so the integrator understands the full scope before running git merge. Accepts a target branch (default: main).
- ▌ Github Issues Setup · dazzlemlInitialize standard GitHub issues (#1 Roadmap, #2 Quick Notes) with the canonical label set for a new project.
- ▌ Dev Workflow Process · dazzlemlInitiates a structured analysis of a problem using the DEV WORKFLOW PROCESS with specific contextual guidance.