MonsieurBarti

Use when reviewing architecture decisions. C4 model, dependency inversion, hexagonal boundaries.

Use during discuss-slice and plan-slice to stress-test a spec or plan against domain vocabulary before approval.

Use during research-slice to build disposable experiments that answer technical uncertainty.

Use during project:init to scaffold repo-level agent configuration (CLAUDE.md, AGENTS.md, tracker labels).

Use when debugging. 4-phase investigation, root cause, minimal fix.

Use when implementing features/fixes. Iron law: ¬∃ production code without failing test.

Use during discuss-slice to synthesize conversation context into a formal PRD with user stories, implementation decisions, testing decisions, and exclusions.

Use ad-hoc to classify the current slice or task into type and readiness state using the tff state model.

Use when making git commits. Conventional commit format and rules.

Use when running security review on a diff/PR/file. Research-vs-Reporting discipline, three-tier confidence gate, attacker-controllable taxonomy.

Use when invoking plannotator for approval review of generated .tff-cc milestone artifacts (plan, verification, spec).

Use when starting design or discovery work. MUST use before any creative work.

Use when the user wants to report a bug, file an issue, or send feedback about tff-cc. Triggers on phrases like 'report issue', 'file a bug', 'tff-cc broke', 'something is wrong with tff-cc', 'open an issue', '/report-issue'.

Use when creating, refining, or composing skills. Evidence-driven pattern analysis.

Use when verifying acceptance criteria. Binary verdict per criterion. Evidence-based.

Use when creating or modifying agents. Standardized template, identity-only, skills-loaded.

Produce a multi-tab, fact-checked HTML architecture document for a codebase by orchestrating lumen-diagram, lumen-guide, and lumen-fact-check in a fixed sequence. Invoke when user asks for an architecture doc, design doc, or to "document this architecture".

Produce a magazine-quality scroll-snap HTML deck announcing a new feature, release, or milestone. Combines a state recap, metrics charts, a hero diagram, and an assembled slide deck. Invoke for "launch deck", "release deck", "announcement slides".

Produce a multi-tab HTML postmortem for a production incident. Combines a scoped recap, a mermaid sequence-diagram timeline, impact charts, and a structured write-up (Summary / Timeline / Impact / Root Cause / Action Items). Invoke for "postmortem", "post-mortem", "incident review".

Produce a multi-tab HTML "project page" combining a project recap, an architecture diagram, 1–2 key-metrics charts, and a narrative guide. Use to showcase a project's current state, architecture, and metrics in one shareable artifact.

Generate a complete documentation set for a project by orchestrating multiple lumen composites (architecture-doc + readme-pack). Higher-altitude than a single composite — produces 2+ HTML artifacts that together rebuild a reader's mental model of the project. Invoke for "document this project", "full docs pack", "doc dump for X".

Bundle release artifacts for a tagged release by orchestrating lumen-launch-deck + lumen-readme-pack and producing a "what shipped" archive. Higher-altitude than a single composite — produces multiple HTML artifacts plus a release index. Invoke for "release pack", "release artifacts", "package the release".

Generate single-file HTML+SVG or interactive Chart.js data chart (bar, pie, line, area, scatter, radar, funnel, bubble, table, interactive-bar, interactive-line). Invoke when user provides numeric data and asks for chart, graph, plot, trend, distribution, proportion, or comparison.

Generate single-file HTML diagram (architecture, flow, sequence, ER, state, gantt, network, dependency graph, AI pattern). Invoke when user asks to draw, diagram, visualize, sketch architecture, show flow, or describes system / pipeline / topology to illustrate.

Adds a new domain error to an existing module — extends the module base error, sets error code and monitoring flag, and updates the exception filter mapping. Use when a command or query needs to throw a new business error. Requires /api-add-module.

Adds a new HTTP endpoint to an existing controller — request/response DTOs, Swagger decorators, and command/query wiring. Use after the module, entity, and command/query already exist. Requires /api-setup-shared and /api-add-module.

Creates a CQRS command + handler + in-memory test in a NestJS hexagonal module. Use when adding a write operation (state change) to a module. Command and handler live in the same file. Requires /api-setup-shared.

Creates a complete domain entity in a NestJS hexagonal module — entity, repository interface, test builder, unit tests, SQL repository, in-memory repository, mapper, and optionally domain events. Supports both standalone entities and AggregateRoot entities with domain event publishing. Use when adding a new aggregate to a module's domain layer.

Creates a domain event handler that reacts to events from another module (or the same module). Includes handler class and unit test. Use after domain events have been created with /api-add-domain-entity.

Creates a CQRS query + handler + integration test in a NestJS hexagonal module. Use when adding a read operation (data retrieval) to a module. Query and handler live in the same file. Handlers bypass the domain and repository layers entirely — direct PrismaService only. Requires /api-setup-shared.

Creates all shared infrastructure required for hexagonal NestJS modules — TypedCommand, TypedQuery, TypedCommandBus, TypedQueryBus, CorrelationId decorator, PaginatedQueryBase, BaseDomainError, BaseFeatureExceptionFilter, the BaseLogger pattern (wrapping nestjs-pino), ZodValidationPipe with @ZodSchema decorator, PrismaService + PrismaModule (global), validateEnv helper, CqrsInterceptor (bus observer), LoggingInterceptor (HTTP request logger), DomainEvent base class, SqlRepositoryBase (generic Prisma CRUD + auto event publishing), and IDateProvider + DateProvider + FakeDateProvider (date abstraction for deterministic testing). Run once per project before using other api-* skills.

Bootstraps a brand-new NestJS backend with hexagonal architecture — creates the project via the Nest CLI, removes boilerplate, installs the required stack (Prisma, Vitest, pino, CQRS, Fastify, Biome), configures package.json scripts, sets up git hooks via lefthook, creates the project-specific env config, and generates a .claude/CLAUDE.md project guide. Run once on a fresh directory; follow with /api-setup-shared to complete the shared infrastructure.

Understand an unfamiliar codebase using the knowledge graph. Discovers structure, clusters, key symbols, and execution flows. Use when: "explore" | "understand" | "new codebase" | "orient" | "what does this do"

Review changes before committing — detect which symbols and processes are affected by the current diff. Use when: "pre-commit" | "review changes" | "what breaks" | "what am i changing" | "impact of diff"

Scope a refactor with impact analysis before touching code. Identifies blast radius, dependent processes, and downstream risks. Use when: "refactor" | "rename" | "scope refactor" | "safe refactor" | "impact" | "blast radius"

Biological memory for the agent. Use when: observing facts worth persisting | recalling prior context | reacting to errors | making architectural decisions | resolving conflicts. Triggers: "remember" | "recall" | "i learned" | "we decided" | "the issue was" | "never do X" | "always do Y".

Author path-scoped rule files for pi-rules. Use when: creating/editing .pi/rules/*.md or .claude/rules/*.md | "add a project rule" | "write a pi rule" | designing always-on conventions. Triggers: "rule", "convention", "paths".

Compress runtime output or markdown files. Levels: off|lite|standard|ultra|symbolic. Triggers: "caveman mode" | "ultra compress" | "shorten output" | "compress skill" | "compress this file" | "make it formal" | "use formal notation" | "talk like caveman".