oduffy-delphi — AI Agent Skills on SkillsAuth

Census/launch/teardown lifecycle for a repo's declared app under coordinator.local.md's app_session config -- complementary to the platform's built-in run skill, not competing with it.

tools5

coordinator-update

Check for a published coordinator update and advise a preserve-by-default migration path — never a blind overwrite.

testing5

dogfood

Smoke-driven fix-through loop — invoke a newly-built thing (skill, script, pipeline, install process) and fix bugs until it works or gets replanned. Binary outcome only: converge or switch gears.

devops4

execute-plan

Execute a PM-approved implementation plan directly in the coordinator session

tools4

architecture-audit

Rotational arch audit — scores systems, audits top-priority, packages spinoff candidates. Never edits code; updates Last-targeted-audit clock.

development4

Invoke on any planning trigger from the PM — "plan", "let's plan", "write a plan", "draft a plan", "break this down", "plan the implementation" — for decision-weight work (multi-file, abstraction, cross-system, agent scaffold, reversed prior). Triage lives inside the skill, not in EM pre-skill judgment; writing a plan to disk without invoking this skill is a doctrine violation.

testing4

learn-lessons

Processes state/lessons.md as doctrine change-requests. 3 modes: local, central, recheck. Triggers on triage/trim/process lessons, promote universals.

testing4

spinoff

PM-GATED. Fork a mid-session topic into its own pickup-able handoff (a spinoff). Never EM-initiated.

tools4

validate

Use before commit, /merge-to-main, /workday-complete, or to validate repo state. Resolves and runs the project's configured fast-test command.

testing4

review

Use when a plan / design doc / RFC is ready for review or plan-review findings have landed.

documentation4

percolate

Dry-run then confirm publish of files from a working source tree to a named publish-repo target. Wraps publish.sh with gate + CI smoke.

testing4

project-onboarding

Use when starting in a new project repo, when /update-docs reports tracker_missing, or on first coordinator-plugin run.

tools4

debt-triage

EM-PM conversation to review and prioritize the technical debt backlog. Triggers on demand or when open item count exceeds 20.

tools4

plan

Invoke on any planning trigger from the PM — "plan", "let's plan", "write a plan", "draft a plan", "break this down", "plan the implementation" — for decision-weight work (multi-file, abstraction, cross-system, agent scaffold, reversed prior). Triage lives inside the skill, not in EM pre-skill judgment; writing a plan to disk without invoking this skill is a doctrine violation.

testing4

repo-setup

First-time repo setup (default: single-repo), or `--batch` for fleet setup over working-repos.yaml. Consolidates /project-onboarding and /bootstrap-repos into one surface.

data-ai4

workstream-complete

Wrap up finished work — capture lessons, update docs

documentation4

coordinator-update

Check for a published coordinator update and advise a preserve-by-default migration path — never a blind overwrite.

testing4

percolate

Dry-run then confirm publish of files from a working source tree to a named publish-repo target. Wraps publish.sh with gate + CI smoke.

testing4

codex-review-gate

Run a Codex code review as a second-opinion gate. Returns structured result with findings or graceful skip. Used by /bug-sweep --codex-verify and /workweek-complete Step 7.4.

development4

percolate

Dry-run then confirm publish of files from a working source tree to a named publish-repo target. Wraps publish.sh with gate + CI smoke.

testing4

review-code

Use when a code change / diff / PR is ready for review or code-review findings have landed.

development4

roadmap-planning

PM-GATED. Roadmap shaping from research/deep-dive/peer-repo inputs via Synthesize → Substantiate → Plan → Dispatch pipeline producing kind:spinoff-roadmap stubs with deployment_state, blocks/blocked_by graph, machine-readable STUB-INDEX. Stubs cite PM-approved OVERVIEW + research corpus, not EM hand-waving. Triggers — "shape a roadmap", "plan a sprint sequence from this research".

development4

parallel-code-review

Pre-merge weekly code-review gate — N code-semantics chunk reviewers (Sonnet) + 3 mechanical specialist workers + no-rewrite synthesizer (BLOCKED/WARN/OK). The Staff Engineer runs a separate post-gate architecture pass, not the gate itself. Invoked only from /workweek-complete.

development4

consolidate-git

Cleans up branch sprawl. Triggers: consolidate branches, clean up branches, stale branches, merge all branches.

tools4

dogfood

Smoke-driven fix-through loop — invoke a newly-built thing (skill, script, pipeline, install process) and fix bugs until it works or gets replanned. Binary outcome only: converge or switch gears.

devops4

merging-to-main

Use when a branch is ready to merge to main. Drafts release notes, creates PR, waits for CI, merges, cleans up.

testing4

setup

coordinator-claude install-chain walker (chain step 5 of 5 — DAG root; no upstream deps to walk). Reads the install manifest, observes empty direct_deps, emits the DAG-root terminal banner, and exits 0. Not to be confused with coordinator:install (OSS plugin install of the coordinator package) or coordinator:repo-setup (consumer-project first-time setup of the coordinator integration). Trigger phrases: /coordinator:setup, set up coordinator-claude, run the install chain for coordinator.

tools4

new-project

Create + scaffold + onboard a greenfield repo from any cwd in one command. Creates the dir, git-inits, scaffolds a vendored stack template, delegates coordinator onboarding to repo-setup, optional remote. Sibling of repo-setup (create-new vs onboard-existing).

tools4

plan-delivery-audit

Triangulate plan-claim / code-reality / review oracles to classify each plan into DELIVERED+REVIEWED / DELIVERED-UNREVIEWED / PARTIAL / IN-FLIGHT / ABANDONED. Run after any crash or 'did we actually finish what we think we finished?' moment.

development4

setup

deep-research-claude install-chain walker (chain step 4 of 5). Walk the direct_deps declared in docs/install/agent-install-manifest.json, probe coordinator-claude (soft dep), dispatch the chain-walker subagent, and report. Trigger phrases: /deep-research:setup, set up deep-research, run the install chain for deep-research.

documentation4

codex-review-gate

Run a Codex code review as a second-opinion gate. Returns structured result with findings or graceful skip. Used by /bug-sweep --codex-verify and /workweek-complete Step 7.4.

development4

architecture-audit

Rotational arch audit — scores systems, audits top-priority, packages spinoff candidates. Never edits code; updates Last-targeted-audit clock.

development4

bug-sweep

Systematic codebase bug hunt — find and fix all AI-fixable bugs in-session, defer blocked ones to backlog

development4

coordinator-update

Check for a published coordinator update and advise a preserve-by-default migration path — never a blind overwrite.

testing4

finishing-a-development-branch

Use when implementation is complete and tests pass — presents structured options for merge, PR, or cleanup to integrate the work.

development4

roadmap-planning

PM-GATED. Roadmap shaping from research/deep-dive/peer-repo inputs via Synthesize → Substantiate → Plan → Dispatch pipeline producing kind:spinoff-roadmap stubs with deployment_state, blocks/blocked_by graph, machine-readable STUB-INDEX. Stubs cite PM-approved OVERVIEW + research corpus, not EM hand-waving. Triggers — "shape a roadmap", "plan a sprint sequence from this research".

development4

spinoff

PM-GATED. Fork a mid-session topic into its own pickup-able handoff (a spinoff). Never EM-initiated.

tools4

parallel-code-review

Pre-merge weekly code-review gate — N code-semantics chunk reviewers (Sonnet) + 3 mechanical specialist workers + no-rewrite synthesizer (BLOCKED/WARN/OK). The Staff Engineer runs a separate post-gate architecture pass, not the gate itself. Invoked only from /workweek-complete.

development4

pickup

Resume work from a handoff or action a cross-repo memo — grab the baton and run

tools4

brainstorming

Use for new feature requests, vague/ambiguous requirements, or multi-subsystem decomposition before plan mode.

data-ai4

enrich-and-review

Run enrichment pipeline on chunk directories

devops4

finishing-a-development-branch

Use when implementation is complete and tests pass — presents structured options for merge, PR, or cleanup to integrate the work.

development4

staff-session

PM-GATED: ask first; never from subagent. Agent Teams collaborative planning/review for architectural decisions only. Modes: plan, review.

testing4

systematic-debugging

Root-cause discipline for ONE identified bug, test failure, or unexpected behavior — pin the premise, reproduce, trace to source, fix at source, verify. For a single known issue, not a codebase sweep.

development4

review-code

Use when a code change / diff / PR is ready for review or code-review findings have landed.

development4

execute-plan

Execute a PM-approved implementation plan directly in the coordinator session

tools4

code-health

Night-shift code health review — queries completion entries for today's surfaces, dispatches reviewer, applies findings, updates health tracking.

development4

merging-to-main

Use when a branch is ready to merge to main. Drafts release notes, creates PR, waits for CI, merges, cleans up.

testing4

debt-triage

EM-PM conversation to review and prioritize the technical debt backlog. Triggers on demand or when open item count exceeds 20.

tools4

review-code

Use when a code change / diff / PR is ready for review or code-review findings have landed.

development4

staff-session

PM-GATED: ask first; never from subagent. Agent Teams collaborative planning/review for architectural decisions only. Modes: plan, review.

testing4

handoff

Involuntary mid-workstream save-state under context pressure. By definition a continuation, never a workstream ending — see Step 0.

tools4

pickup

Resume work from a handoff or action a cross-repo memo — grab the baton and run

tools4

repo-setup

First-time repo setup (default: single-repo), or `--batch` for fleet setup over working-repos.yaml. Consolidates /project-onboarding and /bootstrap-repos into one surface.

data-ai4

bug-sweep

Systematic codebase bug hunt — find and fix all AI-fixable bugs in-session, defer blocked ones to backlog

development4

brainstorming

Use for new feature requests, vague/ambiguous requirements, or multi-subsystem decomposition before plan mode.

data-ai4

workstream-start

Orient session — preflight, load context, choose work

tools4

shape

Converge with the PM on the SHAPE of a problem (the PRD half) before any solutioning. Strategic, not tactical. Exit gate chains into coordinator:plan.

data-ai4

learn-lessons

Processes state/lessons.md as doctrine change-requests. 3 modes: local, central, recheck. Triggers on triage/trim/process lessons, promote universals.

testing4

code-health

Night-shift code health review — queries completion entries for today's surfaces, dispatches reviewer, applies findings, updates health tracking.

development4

doctor

coordinator-claude doctor (stub — install-chain contract conformance; full doctor skill is follow-on work per docs/plans/2026-06-15-coordinator-install-chain-application-phase-b.md §13)

documentation4

review

Use when a plan / design doc / RFC is ready for review or plan-review findings have landed.

documentation4

workstream-start

Orient session — preflight, load context, choose work

tools4

plan-delivery-audit

Triangulate plan-claim / code-reality / review oracles to classify each plan into DELIVERED+REVIEWED / DELIVERED-UNREVIEWED / PARTIAL / IN-FLIGHT / ABANDONED. Run after any crash or 'did we actually finish what we think we finished?' moment.

development4

workstream-complete

Wrap up finished work — capture lessons, update docs

documentation4

handoff

Involuntary mid-workstream save-state under context pressure. By definition a continuation, never a workstream ending — see Step 0.

tools4

validate

Use before commit, /merge-to-main, /workday-complete, or to validate repo state. Resolves and runs the project's configured fast-test command.

testing3

systematic-debugging

Root-cause discipline for ONE identified bug, test failure, or unexpected behavior — pin the premise, reproduce, trace to source, fix at source, verify. For a single known issue, not a codebase sweep.

development3

session-start

Orient session — preflight, load context, choose work

tools3

enrich-and-review

Run enrichment pipeline on chunk directories

devops3

session-end

Wrap up finished work — capture lessons, update docs

documentation3

plugins/coordinator/skills/daily-review

Strategic daily review — inventory today's work, summarize what shipped, get architectural perspective

tools2

lessons-trim

DEPRECATED ALIAS — invokes coordinator:lesson-triage in project-local mode. Periodic maintenance of lessons files (trim stale entries, merge duplicates, clean feature-scoped files). Invoked by /update-docs (Phase 6) or standalone. Removal due 2026-05-26 (one cadence cycle after lesson-triage ship).

documentation2

plugins/coordinator/skills/generate-repomap

Generate a ranked repository map for LLM context injection

tools2

plugins/coordinator/skills/review-dispatch

Route artifacts to the right reviewer

tools2

requesting-staff-session

Use when the PM asks for staff input on a plan or review, when the EM needs multi-perspective planning or critique, or when deciding between /staff-session and /review-dispatch. Guides tier selection, team composition, and scoping.

testing2

architecture-rotation

Run the weekly architecture audit rotation — score systems, audit the highest-priority target, apply findings, and update the health ledger

testing2

shape

Converge with the PM on the SHAPE of a problem (the PRD half) before any solutioning. Strategic, not tactical. Exit gate chains into coordinator:plan.

data-ai2

writing-plans

This skill should be used when requirements are clear and the task needs decomposition into executable chunks — before touching code. Triggers on: 'write a plan', 'break this down', 'plan the implementation'.

development2

verification-before-completion

Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always

data-ai2

consolidate-git

Cleans up branch sprawl. Triggers: consolidate branches, clean up branches, stale branches, merge all branches.

tools2

using-git-worktrees

Use when starting feature work that needs true branch-level isolation (separate PRs, different base branches, long-lived parallel features) - NOT for avoiding file conflicts during parallel agent dispatch (use sequential execution instead)

data-ai2

requesting-code-review

Use when completing tasks, implementing major features, or before merging to verify work meets requirements

development2

using-git-worktrees

Use when starting feature work that needs true branch-level isolation (separate PRs, different base branches, long-lived parallel features) - NOT for avoiding file conflicts during parallel agent dispatch (use sequential execution instead)

data-ai1

dispatching-parallel-agents

Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies

testing1

skill-discovery

This skill should be used when unsure which skill or command applies to a task, or when the user asks 'what skills exist', 'what commands are available', or 'help me find the right tool'.

tools1

consolidate-git

Use when the repo has multiple stale branches that need cleaning up — inventories all branches, absorbs unique commits into the current branch, deletes stale branches, and merges to main. This skill should be used when the user asks to "clean up branches", "consolidate branches", "consolidate git", "merge all branches", or mentions stale/old branches that need cleanup.

data-ai1

brainstorming

Use when the PM asks for a new feature or capability, or you're about to enter plan mode without a clear spec. Also use when requirements are vague, ambiguous, or span multiple subsystems that need decomposition.

data-ai1

tracker-maintenance

Maintain the unified project tracker at docs/project-tracker.md — marks completion, archives shipped work, updates dependencies, and sweeps for untracked commits. This skill should be used when the user asks to clean up or update the project tracker, archive completed work, check for untracked commits, or resolve stale dependencies. Invoked by /update-docs (Phase 5) or standalone.

testing1

remember

Quick-write a lightweight session note to memory/sessions/note.md. For full session wrap-up, use coordinator:session-end or /handoff.

documentation1

codex-review-gate

Run a Codex code review as a second-opinion gate. Returns structured result with findings or graceful skip. Used by /bug-sweep and /workday-complete.

development1

tracker-maintenance

Maintain the unified project tracker at docs/project-tracker.md — marks completion, archives shipped work, updates dependencies, and sweeps for untracked commits. This skill should be used when the user asks to clean up or update the project tracker, archive completed work, check for untracked commits, or resolve stale dependencies. Invoked by /update-docs (Phase 5) or standalone.

testing1

artifact-consolidation

Use when artifact directories are bloated, during periodic maintenance, or when disk usage from session debris is excessive. Prunes and consolidates accumulated session artifacts — plans/, archive/handoffs/, stale task dirs. Supports dry-run mode. Standalone invocation only — not part of /update-docs.

documentation1

artifact-consolidation

Use when artifact directories are bloated, during periodic maintenance, or when disk usage from session debris is excessive. Prunes and consolidates accumulated session artifacts — plans/, archive/handoffs/, stale task dirs. Supports dry-run mode. Standalone invocation only — not part of /update-docs.

documentation1

brainstorming

Use when the PM asks for a new feature or capability, or you're about to enter plan mode without a clear spec. Also use when requirements are vague, ambiguous, or span multiple subsystems that need decomposition.

data-ai1

merging-to-main

Use when work on a branch is ready to merge to main — creates PR, waits for CI, merges, cleans up.

testing1

test-driven-development

Use when implementing any feature or bugfix, before writing implementation code

development1

review-blueprint

LLM-powered Blueprint review with RAG grounding. Entry point for the /coordinator:review-blueprint command pipeline. Sub-skill rag-fetch.md handles domain classification and RAG retrieval.

devops1

systematic-debugging

This skill should be used when encountering any bug, test failure, or unexpected behavior — before proposing fixes. Triggers on: 'something is broken', 'test is failing', 'unexpected behavior', 'debug this'.

development1

stuck-detection

This skill should be used when detecting repetitive actions, oscillating between approaches, or stalling without progress — the three stuck patterns. Referenced by agent prompts for self-monitoring.

testing1

atlas-integrity-check

Check today's changed files against the architecture atlas — flags narrative-drift when changes touch systems the atlas description doesn't mention or reflect. On RAG repos: narrative-drift findings. On non-RAG repos: also flags unmapped files as potential new systems. Invoked by /update-docs (Phase 11) or standalone.

testing1

stuck-detection

This skill should be used when detecting repetitive actions, oscillating between approaches, or stalling without progress — the three stuck patterns. Referenced by agent prompts for self-monitoring.

testing1

skill-discovery

This skill should be used when unsure which skill or command applies to a task, or when the user asks 'what skills exist', 'what commands are available', or 'help me find the right tool'.

tools1

finishing-a-development-branch

Use when implementation is complete, all tests pass, and you need to decide how to integrate the work - guides completion of development work by presenting structured options for merge, PR, or cleanup

development1

atlas-integrity-check

Check today's changed files against the architecture atlas file-index.md — flags unmapped files as potential new systems. This skill should be used when verifying that new or changed files are mapped in the architecture atlas, or after adding new modules or directories. Invoked by /update-docs (Phase 11) or standalone.

testing1

test-driven-development

Use when implementing any feature or bugfix, before writing implementation code

development1

requesting-staff-session

Use when the PM asks for staff input on a plan or review, when the EM needs multi-perspective planning or critique, or when deciding between /staff-session and /review-dispatch. Guides tier selection, team composition, and scoping.

testing1

writing-skills

Use when creating new skills, editing existing skills, or verifying skills work before deployment

testing1

receiving-code-review

Use when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation

development1

project-onboarding

Use when starting work in a new project repository, when /update-docs reports tracker_missing, or when a marketplace user runs the coordinator plugin for the first time.

tools1

requesting-code-review

Use when completing tasks, implementing major features, or before merging to verify work meets requirements

development1

remember

Quick-write a lightweight session note to memory/sessions/note.md. For full session wrap-up, use coordinator:session-end or /handoff.

documentation1

handoff-archival

Archive consumed handoffs — moves superseded or PM-approved handoffs from tasks/handoffs/ to archive/handoffs/. Invoked by /update-docs (Phase 8) or standalone. Does NOT auto-archive based on age alone.

documentation1

codex-review-gate

Run a Codex code review as a second-opinion gate. Returns structured result with findings or graceful skip. Used by /bug-sweep and /workday-complete.

development1

writing-skills

Use when creating new skills, editing existing skills, or verifying skills work before deployment

testing1

receiving-code-review

Use when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation

development1

lessons-trim

Periodic maintenance of lessons files — trims stale entries, merges duplicates, and deletes exhausted feature-scoped files. Implements the 'Periodic trim' rule from CLAUDE.md's Self-Improvement Loop. This skill should be used when lessons.md is getting long, when a feature is complete and its lessons file should be cleaned up, or when periodic housekeeping is needed. Invoked by /update-docs (Phase 6) or standalone.

documentation1

dispatching-parallel-agents

Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies

testing1

writing-plans

This skill should be used when requirements are clear and the task needs decomposition into executable chunks — before touching code. Triggers on: 'write a plan', 'break this down', 'plan the implementation'.

development1

handoff-archival

Archive consumed handoffs — moves superseded or PM-approved handoffs from tasks/handoffs/ to archive/handoffs/. Invoked by /update-docs (Phase 8) or standalone. Does NOT auto-archive based on age alone.

documentation1

debt-triage

Use when the technical debt backlog needs review and prioritization, on demand, or when the backlog exceeds 20 open items. This is an EM-PM conversation, not a dispatched agent — the EM reads the backlog, applies judgment, and presents recommendations. Note: weekly-architecture-audit will increasingly insist as the count grows — mild concern at >20, visible disappointment at >30, and a full coffee-down stare-down at >40.

testing1

verification-before-completion

Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always

data-ai1

validate

Use when about to commit, before /merge-to-main, before /workday-complete, or when the user asks to validate the repo state. Runs all CI validation checks locally.

testing1