ss

Plugin ss from marketplace specswarm-marketplace in MartyBonacci/specswarm. Commands read from the repo's manifest — 1 month ago.

Skills (10)

• ss-build

  Spec-driven feature development workflow. Triggers on build/create/add/implement/develop intent.

• ss-fix

  Fix bugs with regression testing and auto-retry. Triggers on fix/debug/broken/error/not-working intent.

• ss-init

  Initialize SpecSwarm for a project with constitution and standards. Triggers on setup/initialize/configure intent.

• ss-metrics

  Show feature metrics and sprint analytics. Triggers on metrics/stats/analytics/how-did-we-do intent.

• ss-modify

  Modify features with impact analysis and compatibility checks. Triggers on modify/change/update/refactor intent.

• ss-release

  Prepare releases with changelog, version bump, and tagging. Confirms on release/tag/publish/version intent.

• ss-rollback

  Safely rollback failed features (DESTRUCTIVE). ALWAYS confirms on rollback/undo/revert/abort intent.

• ss-ship

  Quality-gated merge to parent branch (DESTRUCTIVE). ALWAYS confirms on ship/deploy/merge/release intent.

• ss-status

  Check build progress and session status. Triggers on progress/status/done/how's-the-build intent.

• ss-upgrade

  Upgrade deps/frameworks with breaking change analysis. Triggers on upgrade/update/migrate/modernize intent.

Slash commands (34)

• /ss:analyze-quality

  Comprehensive codebase quality analysis with prioritized recommendations

• /ss:bakeoff

  Calibration loop for taste-heavy choices (v7.16.0). Generates N candidate implementations/outputs, renders a side-by-side contact sheet, captures your verdict, pins the winner with a regression test, and distills the ruling into the taste model. Converts fuzzy judgment into deterministic constants.

• /ss:bugfix

  Regression-test-first bug fixing workflow with smart SpecSwarm/SpecTest integration

• /ss:build

  [migrating to /ss:build] Build complete feature from spec to implementation

• /ss:clarify

  Assume-first gap resolution for the current feature spec (v7.13.0). Auto-fills gaps resolvable from the taste model, spec corpus, or codebase precedent — recorded as provenance-cited entries in the spec's ASSUMPTIONS section — and asks only genuine forks, batched. Reviewing assumptions is ~5× faster

• /ss:complete

  Complete feature or bugfix workflow and merge to parent branch

• /ss:conduct

  Mentor→builder headless dispatch loop (v7.18.0). Author a slice prompt in the proven grammar, dispatch it to a headless builder in the target repo (opus-pinned, per-tree locked, group-killed on timeout, fully audited), then run the INDEPENDENT verification protocol — commit-hash check, real-diff rea

• /ss:constitution

  Create or update the project constitution from interactive or provided principle inputs, ensuring all dependent templates stay in sync.

• /ss:decisions

  Pre-batch all strategic decisions for a feature into one AskUserQuestion-driven sheet, BEFORE /ss:tasks runs. Collapses the mid-chunk "wait, decide this" interrupts into a single front-loaded touchpoint. Hybrid deterministic-scan + decision-miner subagent; writes decision-sheet.md and appends a lock

• /ss:dry-run

  Predict what a SpecSwarm chunk's execution would look like WITHOUT running it. Phase-aware — detects which artifacts exist, simulates everything still ahead. Produces a structured `dry-run.md` with anticipated decisions, risk register, out-of-scope guards, memory gaps, touchpoint estimate, predicted

• /ss:fix

  [migrating to /ss:fix] Fix bugs with test-driven approach and automatic retry

• /ss:go

  The single default entry point (v7.17.0). Walks the full AUTO-MAGIC ladder — specify → assume-first clarify → plan → decisions → tasks → preflight → implement (slice gates + verify) → retrospective — pausing ONLY for the batched decision sheet, the assumptions review, sighted gates, and the ship ble

• /ss:hotfix

  Expedited emergency response workflow for critical production issues

• /ss:implement

  Execute the implementation plan by processing and executing all tasks defined in tasks.md

• /ss:init

  Initialize or refresh SpecSwarm guides for a project / sprint

• /ss:intervention

  Capture a "wait, something feels off" moment as a durable memory file. Each intervention becomes training data for future SpecSwarm automation (preflight, spec-mentor, hooks). Today's manual catch → tomorrow's automatic catch.

• /ss:mentor-init

  Scaffold a mentor directory for the mentor→builder operating pattern (v7.18.0). Stamps out MENTOR-KICKOFF.md (the standing role), process/escalation-boundary.md (three-lane human-escalation model), process/PARKING-LOT.md (verbatim idea capture), the conduct dispatch config, and a SessionStart hook t

• /ss:metrics

  [migrating to /ss:metrics] Feature-level metrics and analytics dashboard

• /ss:modify

  [migrating to /ss:modify] Feature modification with impact analysis

• /ss:notify

  Fire a SpecSwarm notification (sound + desktop banner + optional phone vibrate via Termux) with cascading fallbacks. Useful inside long workflows to ping yourself when a chunk completes or a decision is needed.

• /ss:overnight

  Run a SpecSwarm chunk autonomously while you sleep. Combines pre-batched decisions (v7.6.0) + verification queue (v7.4.0) + headless `claude --print` to execute /ss:preflight → /ss:implement → /ss:verify → /ss:retrospective without user input. Default mode is --check (validate readiness). --exec opt

• /ss:plan

  Execute the implementation planning workflow using the plan template to generate design artifacts.

• /ss:preflight

  Run deterministic preflight checks against a feature's plan.md before /ss:implement. Catches hallucinated versions, missing memory files, spec drift, ambiguous grep patterns, heading typos, and uncited spec assumptions in <5 seconds.

• /ss:release

  [migrating to /ss:release] Comprehensive release preparation workflow

• /ss:retrospective

  Auto-retrospective for a completed SpecSwarm feature/chunk. Gathers feature-scope signals (git log, tasks.md, verify-queue outcomes, captured interventions, MEMORY.md), dispatches the chunk-retrospective subagent, which writes 1-3 durable memory files capturing the chunk's lessons. Solves the "sessi

• /ss:rollback

  [migrating to /ss:rollback] Safely rollback a failed or unwanted feature

• /ss:ship

  [migrating to /ss:ship] Quality-gated merge to parent branch

• /ss:specify

  Create or update the feature specification from a natural language feature description.

• /ss:status

  [migrating to /ss:status] Check status of background sessions and workflows

• /ss:tasks

  Generate an actionable, dependency-ordered tasks.md for the feature based on available design artifacts.

• /ss:upgrade

  [migrating to /ss:upgrade] Upgrade dependencies with breaking change analysis

• /ss:validate

  Run AI-powered interaction flow validation for any software type (webapp, Android app, REST API, desktop GUI)

• /ss:verify

  Adversarial spec-vs-implementation verification for a completed task. Dispatches the spec-mentor subagent with fresh context (no anchoring bias), parses its verdict (PASS / DRIFT / NEEDS-MARTY), and updates the verification queue. The v7.4.0 architectural replacement for a dedicated "mentor session.

• /ss:watchdog

  Background watchdog daemon that monitors a SpecSwarm project for new commits, newly-checked tasks, flagged verifications, and (v7.18.0) mentor→builder conduct runs — pings Marty via ss_notify when something needs attention. Completes the autonomous-execution loop by surviving session restarts and de

Subagents (6)

• chunk-retrospective

  Synthesizes durable memory entries from a completed SpecSwarm chunk's signals — commits, final tasks.md, verify-queue outcomes (PASS/DRIFT), captured interventions, AskUserQuestion answers. Use at /ss:ship time (or just before) to capture the cross-chunk lessons that would otherwise live only in ses

• decision-miner

  Polishes raw decision-candidate signals from a SpecSwarm plan.md into a small, well-formed batch of strategic decisions for AskUserQuestion. Use after `/ss:plan` completes and before `/ss:tasks` runs. Reads plan.md, tech-stack.md, constitution.md, quality-standards.md (locked prior commitments), and

• dry-run-simulator

  Predicts what a SpecSwarm chunk's execution would look like WITHOUT running it. Use after /ss:specify (or any later phase) to surface anticipated decisions, risk register, out-of-scope guards, memory gaps, touchpoint estimate, and predicted artifacts. Reads spec.md plus foundation files plus memory

• specswarm-orchestrator

  Multi-agent orchestration specialist for SpecSwarm. Analyzes tasks.md, identifies parallelizable work streams, routes tasks to specialist agents, and creates integration manifests. Use during implementation phase when --orchestrate flag is set.

• spec-mentor

  Adversarial verification agent for SpecSwarm tasks. Use when a task has just been marked complete and you need a fresh-context, judgment-grade check that the implementation matches the spec verbatim. Compares git diff against §-referenced spec sections + plan.md task block + memory files; returns PA

• task-router

  Routes tasks to specialist agents based on content keyword analysis. Used by the orchestrator for multi-agent builds.

Hooks (9)

• Setup

  ${CLAUDE_PLUGIN_ROOT}/hooks/setup-hook.sh

• SessionStart

  ${CLAUDE_PLUGIN_ROOT}/hooks/orientation-hook.sh

• SessionStart

  ${CLAUDE_PLUGIN_ROOT}/hooks/references-verify.sh

• PostToolUse

  ${CLAUDE_PLUGIN_ROOT}/hooks/quality-check.sh

• PostToolUse

  ${CLAUDE_PLUGIN_ROOT}/hooks/constitution-dispatcher.sh

• PostToolUse

  ${CLAUDE_PLUGIN_ROOT}/hooks/tasks-completion-detector.sh

• Stop

  ${CLAUDE_PLUGIN_ROOT}/hooks/stop-hook.sh

• Stop

  ${CLAUDE_PLUGIN_ROOT}/hooks/go-loop-hook.sh

• Stop

  ${CLAUDE_PLUGIN_ROOT}/hooks/verify-queue-prompt.sh

Similar plugins