Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/lugassawan/swe-workbench/workflowsgit clone --depth 1 https://github.com/lugassawan/swe-workbenchWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00000 | $0.01444 |
| Opus 5 | $0.00000 | $0.00722 |
| Sonnet 5 | $0.00000 | $0.00289 |
| Haiku 4.5 | $0.00000 | $0.00144 |
Grade A, and why
workflows scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 26 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Workflow catalog
All swe-workbench workflow and integration skills available in this plugin. Use the Skill tool to invoke any of these.
swe-workbench:workflow-audit-emit-issues— Post-audit filing: groups codebase audit findings by subsystem (path prefix), discovers issue templates and labels, renders a batch preview, and files one GitHub issue per subsystem onconfirm. Counterpart toswe-workbench:workflow-codebase-audit(read-only); filing mirrorsswe-workbench:workflow-bug-triagePhase 4.swe-workbench:workflow-bug-triage— Investigate-and-file-issue counterpart to /swe-workbench:debug. Iron Law (no fix without root cause), 4-phase loop, files structured issue with code-path table and impact assessment.swe-workbench:workflow-cleanup-merged— Post-merge cleanup: fast-forward main (which auto-cleans via rimba post-merge hook when active), then verify; falls back torimba removeorgit worktreeshell path; deletes local + remote branch; includes a PR-scoped residual sweep to clean up artifacts left by interrupted prior flows.swe-workbench:workflow-codebase-audit— Cold-start, time-boxed, multi-axis audit sweep with ranked findings, reasoning chains, and counter-evidence calibration.swe-workbench:workflow-codebase-knowledge— Understanding-oriented codebase presentation: architecture overview, module map, public API surfaces, and patterns — read-only, no defects, no new docs.swe-workbench:workflow-commit-and-pr— Pre-merge half: enforces [type] commit format, branch-naming, [no ci] for docs, draft/ready prompt, PR template detection, and post-create /swe-workbench:review CTA.swe-workbench:workflow-development— Full development lifecycle: Branch → Implement → Verify → Review → Deliver. Phase 1 usesrimba addwhen rimba is available; falls back tosuperpowers:using-git-worktrees.swe-workbench:workflow-delegated-implementation— Phase-2 delegation strategy: conditional scope/complexity gate, file-change grouping (Infra/Core/Tests/Wiring axes), dispatch to theswe-workbench:code-implsub-agent, summary-only result consumption, sequential-default with opt-in worktree-isolated parallelism.swe-workbench:workflow-extend— Mid-PR sub-idea capture and implementation onto the existing branch. Skips Phase 1 (Branch), preserves Verify → Review → Deliver, updates the existing PR via workflow-commit-and-pr.swe-workbench:workflow-grill— Grill-me interrogation MODE for a scoped command's clarify step: walks the decision tree one question at a time with recommended answers, self-answers from the codebase, exits on shared understanding or "proceed", hands a Resolved-decisions block back to the command's artifact step. Activated by /swe-workbench:capture /swe-workbench:design /swe-workbench:implement /swe-workbench:architect /swe-workbench:extend /swe-workbench:debug /swe-workbench:hotfix. Not a from-scratch design flow; produces no design doc.swe-workbench:workflow-hotfix— Branch-based P0 hotfix lifecycle: skips the worktree provider (plain branch, never a worktree), implements with TDD relaxed, opens the PR ready-for-review BEFORE verify/review, tags the PR body with a deferred-verification marker, then reconciles it once the regression test is backfilled and verification is green. Invoked by/swe-workbench:hotfix.swe-workbench:workflow-address-feedback— PR-owner feedback loop: fetch open review threads, per-thread ADDRESSED/CLARIFIED/DEFERRED triage, Edit-tool fixes, workflow-commit-and-pr commit, REST reply posting, and GraphQL resolveReviewThread. Invoked by/swe-workbench:address-feedback.swe-workbench:workflow-pr-review— Remote-PR review orchestration, first-pass or followup re-check (selected byMODE): ephemeral worktree + reviewer agent + GraphQL thread dedup + REST inline-comment post + APPROVE/COMMENT submit. Invoked by/swe-workbench:reviewPR mode (first-pass) and/swe-workbench:review --check-followup <N>(followup).swe-workbench:workflow-pr-review-post— Shared posting core: takes a normalized (findings[], decision, byline) payload, deduplicates against existing review threads, posts inline comments (or a batched PR-level comment for findings with no File:Line anchor), applies the self-review gate + diff-scoping flip, submits APPROVE/COMMENT, and offers the address-feedback CTA. Invoked byswe-workbench:workflow-pr-review(both modes) and the/swe-workbench:reviewspecialist PR-mode sub-flow.swe-workbench:workflow-worktree-session— Start, switch, or end a worktree-bound session viaEnterWorktree/ExitWorktree. No claude restart.swe-workbench:ticket-context— Fetch structured context from Jira, Linear, Confluence, and GitHub issues/PRs before starting work.swe-workbench:observability-context— Fetch production observability signals from Sentry before starting work. Framing input only — never a substitute for a captured profile.swe-workbench:comms-context— Fetch on-call and discussion context from Slack threads and PagerDuty incidents before starting work.swe-workbench:workflow-performance-investigation— Profile-first performance investigation runbook: frame+baseline → profile (per-ecosystem tooling matrix) → hand profile to performance-tuner for ranked hotspots → bottleneck taxonomy → one isolated change → before/after measurement → regression guard. Composes principle-performance.swe-workbench:workflow-dependency-upgrade— Dependency-upgrade lifecycle runbook: triage+batch (patch/minor/major, automated vs manual) → bump & regen lockfile → build/test → breakage-triage taxonomy → PR hygiene. Per-ecosystem command matrix; composes principle-security.swe-workbench:workflow-branch-sync— Guided branch sync: delegates the mechanical merge/rebase onto the default branch to rimba (or git), then walks conflicts file-by-file with aswe-workbench:conflict-resolverrecommendation and rationale before applying viaswe-workbench-apply-conflict-resolution. Never auto-pushes. Invoked by/swe-workbench:sync.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 26 lines · 0 tokens per session scan A bfea01753f40
workflows is an agent published in the GitHub repository lugassawan/swe-workbench (2 stars, last pushed 2d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,444 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other agents, from other repositories
architect-review
Master software architect specializing in modern architecture patterns, clean architecture, microservices, event-driven systems, and DDD. Reviews system designs and code changes for architectural integrity, scalability, and maintainability. Use PROACTIVELY for architectural decisions.
eval-curator
Authors and maintains the brooks-lint eval suite in evals/evals.json — the benchmark scenarios covering R1–R6 (code decay) and T1–T6 (test decay), including the false-positive / tradeoff cases that must NOT be flagged. Ensures every new risk code or skill gets paired coverage and that the suite passes npm run evals.…
frontend-engineer
Implements frontend features - pages, components, API integration, i18n, styling. Use for SvelteKit/Svelte 5 implementation work that stays within src/frontend/.
ux-designer
For website changes, owns the "spec" phase as a design spec — information architecture, wireframe/description of the change, interaction notes. Gate 1 approves design, not prose. Persona frontend-ux.
php-reviewer
PHP 8.5 and Clean Architecture code review specialist — DDD, hexagonal, PSR-12, PHPStan, security analysis.
react-reviewer
React 19.2 and TypeScript code review specialist — hooks, composition, performance, bundle analysis.