tomcounsell

claude-standards

Audit Claude Code assets (skills, commands, subagents, hooks, MCP) against documented standards, optionally applying conformance fixes. Use when reviewing Claude Code hygiene; invoke explicitly.

do-skills-audit

Audit skill quality: lint structure, descriptions, rot, orphans; --arch for architecture dispositions. Use when auditing, linting, or checking skills.

skillify

Capture this session's repeatable process into a reusable skill. Triggered by '/skillify', 'capture this as a skill', 'turn this into a skill', 'save this workflow', 'make this repeatable'.

analyze

Strategic analysis of a decision, idea, or plan. Triggered by 'analyze this', 'evaluate/assess this', 'deep dive', 'strategic read', 'your full take', 'should we X or Y'. NOT quick factual lookups.

deepen

Audit a module's code-level observability: structured logging, metrics, tracing. Triggered by 'deepen this module', 'add logging to', 'instrument this', 'make this debuggable'.

do-oop-audit

Audit Python classes for OOP and data-modeling anti-patterns. Use when reviewing class design, boundaries, or hierarchy, or 'check my classes', 'lint class structure', 'scan for design issues'.

observability

Wire up dashboards, alerts, and health checks for a module or service. Triggered by 'add observability to', 'set up monitoring for', 'create a health check', 'wire up alerts', 'observability pass'.

Use when building with test-driven development discipline. Triggered by 'tdd this', 'write tests first', 'red-green-refactor', 'test-driven', or any request to implement a feature using the TDD cycle.

Data & Documents Listed

officecli

Create, analyze, proofread, and modify Office documents (.docx, .xlsx, .pptx) via officecli. Use when creating, inspecting, checking formatting, finding issues, adding charts, or editing Office files.

prime

Use when onboarding to the Valor AI codebase, understanding the system architecture, or when the user asks 'how does this work'. Comprehensive codebase orientation guide.

pthread

Spawn parallel agents for independent tasks that can run concurrently. Triggered by 'parallelize', 'run these in parallel', or when independent subtasks are detected.

mermaid-render

Render Mermaid or Excalidraw files to verified PNG or SVG assets, including hand-drawn diagrams and batch exports.

pencil-design

Work with .pen files via the Pencil MCP or headless CLI. Triggered by '.pen file', 'pencil mcp', 'mcp__pencil__*', 'pencil cli', 'design system pen'. NOT ordinary frontend/CSS work.

improve-preflight

Decide read-only whether an improvement experiment can reach a scored verdict before a freeze spends anything on it.

improve-research

Run a dispatched improvement research case: read the brief, investigate through the recorded kinds, and propose one hypothesis through the improvement CLI.

calendar-sync

Reconstruct a day of repository work from git commits and sync inferred work blocks to its mapped Google Calendar.

do-debrief

Construct a concise spoken executive brief from current work or notes, with decisions and recommendations first.

do-deploy-example

Develop a project-specific deployment runbook or adapt a deployment example to a named hosting environment.

do-integration-audit

Trace whether a feature is reachable and correctly wired across implementation, entrypoints, configuration, tests, and docs.

do-pr-review

Review a pull request against its intent and current diff, with verified actionable findings and an evidence-based verdict.

do-presentation

Create or revise a presentation from research, notes, or code, using audience-appropriate structure and verified visual output.

do-voice-recording

Synthesize supplied text into an audio file for narration, voiceovers, or reading aloud.

authenticity-pass

Check a social draft for grounded metrics, honest constraints, and a defensible point of view before publication.

present

Turn a topic or current conversation into a crafted single-page HTML explainer with useful diagrams.

de-slop

Edit a finished external-facing draft for hollow content, formulaic AI prose, and readability while preserving facts and voice.

build-agent

Implement, evaluate, and launch an agent from a build-sheet using the selected Codex, OpenAI API, or Claude Managed Agents runtime.

computer-use

Inspect and operate native desktop apps, windows, and accessibility controls when the task requires UI interaction.

do-build

Implement a development plan through working code, relevant tests, review, documentation, and a reviewable change.

do-design-audit

Review an existing interface for hierarchy, usability, consistency, accessibility, responsiveness, and visual quality.

do-design-system

Evolve a design-system charter, tokens, and reusable components while keeping design sources and downstream CSS consistent.

do-discover-paths

Explore an authorized browser flow and record a replayable JSON trace with stable selectors and final-state assertions.

do-investigation-issue

Capture an unverified anomaly or integration gap as a GitHub investigation issue with evidence and next experiments.

do-issue

Turn a development request or confirmed defect into a researched, actionable GitHub issue with acceptance criteria.

do-patch

Diagnose and fix a test failure, regression, or review blocker with a focused change and causal verification.

do-plan-critique

Review a development plan for feasibility, missing requirements, complexity, failure modes, and verifiable acceptance criteria.

do-plan

Research and write an implementable development plan with explicit scope, acceptance criteria, dependencies, and verification.

do-sdlc

Advance or complete a development workflow from its current issue, plan, build, test, review, documentation, or merge state.

do-test

Run and interpret meaningful tests for a change, plan, or failure using the repository’s actual test environment.

email

Read, search, draft, or send email using available mail tools, with thread-aware replies and verified delivery status.

google-workspace

Work with Google Drive, Docs, Sheets, Calendar, and Gmail through authenticated Workspace CLI or available connectors.

grill-me

Pressure-test a plan, belief, or strategy through a focused Socratic conversation when the user asks to challenge their thinking.

imagine-agent

Turn a client outcome into an evidence-based agent build-sheet, including scope, evaluation criteria, delivery, and runtime.

new-audit-skill

Create a focused Codex audit skill with observable checks, verified findings, severity definitions, and optional mechanical fixes.

ontologies

Define or reconcile domain terms and bounded contexts in ONTOLOGIES.md when naming or meaning is ambiguous.

weekly-review

Summarize a week, sprint, month, or requested period of engineering work from git history with contributor statistics.

zoom-out

Reassess an ongoing task when priorities drift or repeated fixes fail; recommend the next action against the original goal.

checking-system-logs

Investigate Valor bridge, worker, session, timeout, or error behavior using structured events and service logs.

do-deploy

Verify and deploy merged Valor changes through the established local and fleet update process.

ask-me

Interview the user to resolve decisions left open by research or implementation; ask one contextual question at a time.

Read LinkedIn posts or DMs, draft content, and perform specifically requested posting or engagement actions.

do-docs

Update repository documentation to match a code change, PR, commit, or new behavior without duplicating sources of truth.

do-merge

Merge a requested pull request after verifying current-head approval, checks, mergeability, and repository gates.

Web & Frontend Listed

frontend-design

Build or refine frontend interfaces with deliberate visual hierarchy, coherent styling, accessible interaction, and responsive behavior.

reclassify

Change a plan classification among bug, feature, and chore while it is still in Planning status.

cowork

Create or maintain recurring agent workflows, Codex automations, or an explicitly requested Claude cloud routine.

pen-design

Create or edit .pen designs with available Pen tools or the pen.dev CLI, preserving design files and editor state.

ebook-ingest

Convert and clean legally available ebooks or supplied EPUB, PDF, or MOBI files into Markdown and optional retrieval chunks.