Command
Help the PR owner address review feedback — fetches open threads, presents a per-thread triage (ADDRESSED / CLARIFIED / DEFERRED), applies fixes, commits, posts per-thread replies, and resolves review threads.
Command
Author an ADR, RFC, or cross-service contract via the architect subagent.
Command
Cold-start, time-boxed, multi-domain audit sweep — surfaces ranked findings with reasoning chains across security, performance, reliability, tooling, and testing.
Command
Capture an idea, improvement, or bug as a well-framed GitHub issue (works in any repo).
Command
Clean up worktree and local + remote branch for a merged PR.
Command
Present a structured knowledge document for this codebase — architecture overview, module map, public API surfaces, and patterns. Knowledge-presentation only (no defect ranking). Use /swe-workbench:audit-codebase for defect sweeps; /swe-workbench:document for generating new prose docs.
Command
Run an unattended review→fix→re-review loop on the current branch until the reviewer has no findings at or above Medium, then push once. Findings stay in a local scratch file and are never posted to GitHub. Stops early on oscillation, a red test suite, or a 4-review cap.
Command
Diagnose a bug, failing test, or unexpected behavior — root-cause first, minimal fix, regression test. Performance-shaped symptoms (slow, latency, profiling) are routed to workflow-performance-investigation via one clarifying question.
Command
Consult the senior-engineer subagent for an architectural decision.
Command
Read-only preflight check of runtime dependencies (gh, git, jq, rimba, claude, python3) plus gh auth status and .claude/cache/ repo hygiene. Prints a green/red table; never modifies state. Exit 0 regardless of findings; exits 1 only if the swe-workbench-doctor command is not on PATH.
Command
Generate or update documentation (README, ADR, ARCHITECTURE, inline comments) via the tech-writer subagent.
Command
Capture a mid-PR sub-idea and implement it onto the current open PR's branch — no new branch, no new PR. Preserves Verify → Review → Deliver.
Command
Hand work to Claude Code or Pi through a bounded semantic checkpoint, resume or recover a checkpoint, inspect status, or close a completed handoff.
Command
Ship a fast branch-based P0 hotfix — diagnose, implement, open a ready-for-review PR, then verify + review after (deferred by design).
Command
Drive a feature implementation end-to-end from a ticket or spec — branch, plan, build, verify, review, deliver.
Command
Invoke the migrator subagent to plan and execute a multi-deployment migration via expand → backfill → dual-write → switch → contract.
Command
Invoke the refactorer subagent for a behavior-preserving refactor.
Command
File a plugin bug or feature request directly into lugassawan/swe-workbench from any directory.
Command
Review the current git diff — auditor selected by --mode (general, security, a11y, deps, perf, tests, contributor-trust, ux) or auto-inferred from the diff when omitted. Pass a PR number to review a specific PR; use --check-followup to re-check a PR after the owner has addressed feedback.
Command
Depth-first security audit of the current git diff — OWASP Top 10, secret leakage, insecure-by-default APIs, and language-specific foot-guns. Pass a PR number to audit a specific PR's diff.
Command
Bring the current branch up to date with the default branch — delegates the mechanical merge/rebase to rimba (or git), then walks through any conflicts file-by-file with a recommended resolution before applying. Never auto-pushes. Pass --rebase to rebase instead of merge. Pass --check-redundancy to also surface…
Command
Invoke the test-writer subagent to add focused, behavioural tests in the target language's idiom. Pass --mode e2e (or --e2e) to run the browser-driven E2E pipeline instead, or --mode e2e-live for an ephemeral, human-watched browser walkthrough with no spec file written.
Hook
Runs before the agent uses a tool for Bash, Skill, Read, Edit, Write, Write, Edit, Bash, Edit and Write tool calls, executing bash_guard.sh, skill_usage_record.sh, worktree_permission_grant.sh, secret_guard.py and handoff_guard.py via bash (5 commands). From lugassawan/swe-workbench.
Hook
Runs after a tool call finishes for Read, Edit and Write tool calls, executing skill_autoload_hint.sh via bash. From lugassawan/swe-workbench.