bork

Plugin bork from marketplace skills in borkweb/skills. Commands read from the repo's manifest — today.

Skills (31)

• agents-md-lint

  Audit and trim AI agent instruction files (AGENTS.md, CLAUDE.md, CONVENTIONS.md, .cursorrules, etc.) by testing which facts an AI agent can discover from code alone. Use when asked to lint, audit, optimize, prune, or trim any agent instruction file in a repository. Removes redundant documentation th

• auto-scope

  Scopes a coding task before implementation — names the handful of files actually relevant (read/edit these), the areas to leave alone, and any open scope questions, so work doesn't start by reading the whole repo. Fires proactively at the start of substantive, multi-file, or unfamiliar-area work; st

• caveman

  Speak in a terse caveman style only when explicitly requested by name. Supports lite, full and ultra; persists until stopped. Ordinary brevity or token-efficiency requests do not activate it.

• complete

  Drive a coding goal to merge-ready through delegated slices, frozen gates and independent review. Use for complete, run the whole loop or take this to merge-ready; merging requires explicit authorization.

• council

  Assess a proposal from multiple perspectives when the user requests a council, debate or adversarial assessment.

• desk

  Start, watch and answer bork `do` runs from a coordinating Claude session (in a herdr tab or over Remote Control); a Codex session inside its sandbox can only wait, show, ack and sometimes answer. Use to start a do run in a repo, wait for runs without polling, see which are done, in progress, blocke

• discernment

  Assess the evidence, uncertainty and failure conditions of a decision or claim when the user requests discernment, calibration or a high-stakes judgment check.

• do

  Run a coding task through the matching pipeline (land an existing change, ship a new goal, or build a plan through complete) with concurrent steps, lean handoffs and council-answered design questions. Use for /do, "fix until safe to land", "keep working until ready" or "build, plan and implement thi

• handoff

  Compact the current conversation into a handoff document for another agent to pick up.

• humanize

  Edit prose for natural, direct language while preserving facts, quotations and voice. Accept inline text or a file; apply file edits unless review-only is requested.

• layman

  Restate existing content in plain English when asked for layman, plain language or no jargon. One-shot unless the user enables sticky mode; never edits files.

• offload

  Make this session the ARCHITECT and offload implementation to a configured builder harness (codex, claude, opencode, pi, grok, or a custom command) you can watch. Reads the session handoff, arbitrates the builder's disagreements, judges raw gate results against frozen criteria plus an independent `r

• pr-review

  Pre-landing PR review that reports instead of merely repairing. Wraps bork:review, adds an independent second pass from the other harness (Claude reviews, Codex seconds it — or the reverse) when the diff needs one, then routes the findings: CRITICAL ones become line-anchored inline PR comments with

• prototype

  Build a throwaway prototype to flesh out a design before committing to it. Routes between two branches — a runnable terminal app for state/business-logic questions, or several radically different UI variations toggleable from one route. Use when the user wants to prototype, sanity-check a data model

• red-pen

  Review or edit prose for clarity, economy and typography when requested. Supports report-only, batched revision or explicitly interactive editing; preserves facts and quotations.

• review-security

  Deep security review of a diff or a specific file, grounded in 20 evidence-based pattern libraries extracted from 400+ real-world bugs in major open-source projects (Linux kernel, OpenSSL, Chromium, Firefox, curl, Go, Rust, Kubernetes, Next.js, etc.). Use when the user says "security review", "secur

• session-budget

  Assess whether compacting or clearing the current session is worth it right now, and whether a reinit-ready handoff that survives /clear would carry real value. Not a threshold nag — it X-rays context composition (how much is stale vs. still load-bearing), which only the agent can see. Use when the

• writing-commits

  Create an authorized git commit or draft its message, grounded in the intended diff and repository conventions. Message-only requests never commit; commit requests do not imply push or merge.

• writing-plan-documents

  Use when writing any plan document or product spec — project plans, PRDs, product specs, proposals, implementation plans, or design docs. Applies ASD-STE100 Simplified Technical English as a writing style layer alongside other planning skills like superpowers:writing-plans.

• writing-sql

  Apply the project’s vertical SQL formatting when authoring or reviewing actual SQL, including migrations and embedded queries. No exception for short queries; generic database discussion does not activate it.

• design-consultation

  Propose a product-specific design system and verified preview, grounded in existing brand assets and user constraints. Covers typography, color, layout, components and accessibility.

• design-review

  Inspect rendered UI for design, responsive and interaction defects against the product’s design system. Apply scoped fixes when requested; report evidence and untested coverage.

• investigate

  Investigate an error or unexpected behavior, establish its cause and apply a verified fix when requested. Uses focused paths for performance, environment and regression problems.

• plan-deep-review

  Review a plan’s strategy and execution risks in expansion, selective expansion, hold-scope or reduction mode. Preserve settled scope and load only relevant domain checks. Use only when the user asks for it, directly or through bork:do.

• plan-design-review

  Review a UI plan for interaction states, design-system fit, accessibility and visual acceptance. Improve the plan when requested; preserve existing brand choices.

• plan-devex-review

  Review developer-facing plans for setup, API/CLI ergonomics, errors and compatibility. Triage covers install and first-run errors; polish and expansion cover broader journeys.

• plan-eng-review

  Review an implementation plan for architecture, contracts, failure paths, concurrency and verification. Improve the plan when requested; do not implement it.

• plan-session

  Run a product or builder design session, resolve consequential unknowns and produce a concrete plan. Reuse supplied context and honor requests to skip questions.

• qa-only

  Perform report-only QA at the selected browser, API, CLI, job or webhook surface with reproducible findings and coverage evidence. May read source for scoping; never edits application files or applies fixes.

• qa

  Test an application at its browser, API, CLI, job or webhook surface and fix reproducible bugs within the requested scope, with evidence and appropriate regression checks. Honors report-only and no-commit requests.

• review

  Review a PR or requested local diff for concrete defects, scope gaps and landing readiness. Report-only unless fixes are requested; includes independent review when warranted.

Slash commands (9)

• /bork:commit

  Create a conventional commit message by analyzing staged changes

• /bork:dream

  Read past do/complete/offload runs out-of-band and propose add, merge, split or retire changes to review-criteria.md, plus per-repo rules and checks for one repo's tooling, each with run citations and counts. Per-row approval; nothing auto-applies. Use when asked to "dream", "consolidate review crit

• /bork:layman

  Restate something in plain, succinct, jargon-free English

• /bork:pre-push

  Critical-only safety review of the diff a push would send. Runs on request, loads the review checklist and reports only CRITICAL findings (SQL injection, auth gaps, race conditions, API contract breaks, weakened gates). It reports and never changes or stops a push.

• /bork:session-budget

  Assess whether to compact/clear this session and prepare a reinit-ready handoff

• /bork:skills-audit

  Audit skills for conflicting instructions, runtime compatibility, invocation, repeated work and selective loading. Recommendations only unless changes are explicitly requested.

• /bork:skills-refine

  Read the back-and-forth that followed a Skill invocation in the current session and propose edits to fold that feedback into the SKILL.md so it handles those cases automatically next time. Per-proposal approval — nothing auto-applies. Use when asked to "refine this skill", "improve this skill", "fol

• /bork:skills-suggest

  Audit recent sessions for repeated tasks and AI-output-then-manual-verification patterns. Produces a ranked list of skill candidates and a drafted SKILL.md for the top pick. Use when asked to "what should I turn into a skill", "suggest skills", "audit my prompts", or run periodically (weekly/monthly

• /bork:status

  Branch status and workflow progress report. Shows current branch state, PR status, what skills have been run, and what's left before the work is shippable. Use when asked to "status", "where am I", "what's left", "progress check", or "am I ready to ship".

Subagents (1)

• triage

  Emergency triage agent for production incidents and urgent bugs. Auto-invoked when the user says "this is broken", "production issue", "urgent fix", "hotfix", "incident", "prod is down", "critical bug", "emergency", "on-fire", or reports errors with urgency language. Routes to the appropriate workfl

Hooks (4)

• SessionStart

  node "${CLAUDE_PLUGIN_ROOT}/skills/core/session-budget/reinit-hook.mjs"

• SessionStart

  node "${CLAUDE_PLUGIN_ROOT}/skills/core/offload/reinit-hook.mjs"

• Stop

  node "${CLAUDE_PLUGIN_ROOT}/skills/core/offload/stop-hook.mjs"

• UserPromptSubmit

  node "${CLAUDE_PLUGIN_ROOT}/skills/core/session-budget/budget-hook.mjs"

Similar plugins