threat-forge

Select the next executable AUTO ThreatForge issue

Shape one ThreatForge issue and populate Project 2 metadata without coding

Compatibility alias for implementing a settled ThreatForge bug issue

Orchestrate one settled ThreatForge issue through verification and preflight

Re-derive and repair the ThreatForge GitHub board after roadmap or code drift

Produce a read-only maintainer briefing from ThreatForge Project 2

Coordinate an explicitly authorized ThreatForge PR lifecycle through owner validation

Run convergent independent review lanes before ThreatForge owner validation

Sweep the whole board and re-triage it — re-derive every non-Done issue's fields, correct drift, split what has outgrown itself, close what no longer deserves to exist, and promote what is now shaped. A long, high-reasoning periodic pass. Use to audit a board that has grown untrustworthy, after a batch of merges, or when a milestone ships.

Self-review ThreatForge changes for AI engineering slop without removing behavior

Run ThreatForge targeted checks and the canonical local verification gate

Self-review a code, test, docs, or copy change for AI-slop before marking it done. Use after implementing any change and before completing a task, PR, issue, or checkpoint. Catches over-engineering, redundant defensive code, comment slop, fake or stub implementations, tautological and weakened tests, hallucinated APIs, duplication, scope drift, and confident-but-wrong logic — without weakening functionality.

Triage one GitHub issue — read it against the codebase, set every field, rewrite the description, link parent and sub-issues, post a justification comment, and promote Backlog to Ready. Use whenever an issue is filed, when an agent files one itself, or when an existing issue's shaping looks stale and needs re-deriving.