Filigree workflow
Skill tachyon-beep/skillpacks/.agents/skills/filigree-workflow
This skill should be used when the user asks to "track work", "create an issue", "find something to work on", "what should I work on next", "triage bugs", "close an issue", "check what's blocked", "plan a milestone", "review sprint progress", "coordinate agents", or when working in a project that uses filigree for issue tracking. Provides workflow patterns, team coordination protocols, and operational guidance for the filigree issue tracker.From its SKILL.md
npx -y skills add tachyon-beep/skillpacks --skill filigree-workflowAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 15 stars15 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
3.6 KB, as published. Nobody here has run it
Filigree Workflow
Filigree is an agent-native issue tracker that stores its data locally in the
project. This skill is procedural knowledge for using it well — as a solo agent
or in a multi-agent swarm. Prefer the mcp__filigree__* tools when available;
fall back to the filigree CLI. filigree --help, filigree <verb> --help
and the MCP tool schemas are the authoritative catalogues.
Core Workflow
filigree session-context → ready / in-progress / critical path
filigree ready → find available work (no blockers)
filigree show <issue-id> → read requirements and context
filigree start-work <issue-id> --assignee <name> → atomically claim + transition into its working status
[do the work, commit code]
filigree close <issue-id> --reason="summary of what was done"
Or skip the search entirely with filigree start-next-work --assignee <name>
to grab the highest-priority startable issue.
Always close with a --reason — it becomes audit trail for the next agent, and
add a comment before any handoff. The next agent has no memory of this
conversation.
Claim Atomically
start-work / start-next-work (CLI) and work_start / work_start_next
(MCP) claim the issue and transition it into its working status in one DB
transaction, with optimistic locking on the assignee. Never chain a claim with
a separate status update — that two-step form races other agents. If someone
else already owns it, the call fails with code: CONFLICT (CLI exit 4); retry
against different work.
Ready ≠ startable. The working status is type-specific (tasks →
in_progress, features →building). Bugs start attriage, which has no single-hop transition into work — they walktriage → confirmed → fixing. So a triage bug is ready but not directly startable:start-workon one returnsINVALID_TRANSITIONnaming the next status, andstart-next-workskips it.readyitems carry astartableflag (and anext_actionhint when false). Pass--advance(CLI) /advance=true(MCP) to walk the soft transitions automatically.
Guidance Sheets
Load these when facing a specific challenge rather than reading upfront:
references/commands.md— the commands and MCP verbs reached for most, the P0–P4 priority scale,--actorattribution, file records and scan findings, cross-product entity bindings, quick decision guidereferences/workflow-patterns.md— triage flows, sprint planning, dependency management, bug lifecycle, event history, archivingreferences/team-coordination.md— multi-agent swarm protocols, handoff conventions, claiming strategies, status update patternsreferences/observations.md— the ambient-note scratchpad: when a note is right, when the thing is task scope instead, triage and promotionreferences/error-codes.md— response envelopes, the completeErrorCodeset, and what to do aboutINVALID_TRANSITION,SCHEMA_MISMATCHandCONFLICTexamples/sprint-plan.json— complete create-plan input template with cross-phase dependencies