filament

Use the filament CLI for project knowledge management, task tracking, lesson capture, and inter-agent coordination. Filament stores entities (tasks, modules, services, agents, plans, docs, lessons), relations between them, messages, and file reservations in a local SQLite database. Use when: managing project context, tracking tasks, recording gotchas and lessons, recording architecture decisions, creating knowledge graph entries, coordinating agent work, or querying project structure. Triggers on: "filament", "add entity", "add task", "add lesson", "track this", "create a task", "gotcha", "lesson", "record a lesson", "search before solving", "capture lesson", "search knowledge", "relate these", "what blocks", "critical path", "ready tasks", "project graph", "knowledge graph", "file reservation", "agent message", "what's next", "dispatch agent", "escalations", "export", "import", "daemon", "tui".

Create an anonymized snapshot of the local fl.db state along with structured bug reproduction steps. The snapshot strips user-identifiable content while preserving schema, entity structure, relations, and the conditions that triggered the bug. Output is a self-contained directory ready to attach to a bug report. Use when: the user says "bug report", "report a bug", "create bug snapshot", "anonymize db", "submit bug", "reproduction steps", or wants to share their local state for debugging. Triggers on: "bug report", "report bug", "bug snapshot", "anonymize", "reproduction steps", "create bug report".

Roleplay a multi-agent orchestration simulation using the filament CLI. Accepts either a custom scenario definition (JSON/YAML format) or falls back to the built-in "web app rewrite" scenario. Exercises entities, relations, tasks, messaging, escalations, reservations, export/import. Triggers on: "start rp", "pause rp", "end rp", "roleplay", "simulation", "simulate agents", "run simulation".

Guide for writing tests following the project's layered test standards. Use when the user says "write tests", "add tests", "test strategy", "what tests to write", "test this module", "where should this test go", or is deciding which test layer to use.

Create an anonymized snapshot of the local fl.db state along with structured bug reproduction steps. The snapshot strips user-identifiable content while preserving schema, entity structure, relations, and the conditions that triggered the bug. Output is a self-contained directory ready to attach to a bug report. Use when: the user says "bug report", "report a bug", "create bug snapshot", "anonymize db", "submit bug", "reproduction steps", or wants to share their local state for debugging. Triggers on: "bug report", "report bug", "bug snapshot", "anonymize", "reproduction steps", "create bug report".

Run manual QA against a CLI binary using a real temp project. Build the release binary, create an isolated test directory, execute commands, capture stdout/stderr/exit codes, and produce a structured results file in the project's .qa/ folder. Use when: the user asks to "manual test", "manual QA", "QA the CLI", "smoke test", "exercise the binary", "end-to-end QA", or wants to verify CLI behavior outside of automated tests. Triggers on: "manual qa", "manual test", "smoke test", "QA", "exercise the CLI", "test the binary", "end-to-end test the CLI".

Recreate bug conditions from a bug report snapshot created by the bug-report-format skill. Loads the anonymized database, replays the reproduction steps, and verifies the bug is reproducible in an isolated environment. Used by maintainers to validate and debug issues. Use when: the user says "recreate bug", "reproduce bug", "load bug report", "replay bug", "verify bug report", or provides a bug report .tar.gz file. Triggers on: "recreate bug", "reproduce bug", "load bug report", "replay bug", "verify bug report", "debug bug report".