ah-engineering

Plugin ah-engineering from marketplace agenthouse-engineering in agenthouse-org/agentic-engineering. Commands read from the repo's manifest — today.

Skills (61)

• ah-architecture

  List, show, check, or write architecture ADR catalogs linked to a work item. Use when maintaining foundational architecture guidelines or linking ADRs before design review.

• ah-assess-market-fit

  Compare market evidence with pre-stated success and stop conditions and recommend proceed, revise, investigate, or stop with an explicit claim level. Use when asking whether an ICP or opportunity is validated, or whether product-market fit evidence supports building.

• ah-assess-story-readiness

  Assess whether one existing work item is complete enough to implement. Use when asking if a story is ready, blocked, or missing facts.

• ah-assess-tech-feasibility

  Assess a proposed technical change, blast radius, encapsulation, rollout options, migration and consumer impact. Use when asking about feasibility or an architecture decision before implementation.

• ah-assessment

  Collect commit-bound feasibility inventory or validate an assessment document bundle. Use when supporting a technical feasibility assessment with reproducible evidence.

• ah-backlog

  Import one markdown or exported JSON work item into .agenthouse/work while keeping its external id. Use when bringing a tracker card or markdown story into local records.

• ah-bundle

  Build an offline install package of this framework and required skills, optionally signed. Use when creating a private distribution or signed update artifact.

• ah-capture-buyer-signal

  Capture a buyer conversation as a single-account signal: goals, objections, and unmet needs, kept separate from market evidence. Use when sales or customer-success staff relay prospect feedback for market validation.

• ah-check-commit

  Check the behavior affected by a commit or change range against its intended outcome. Use when verifying a fix, looking for regressions, or inspecting what a commit changed.

• ah-coding-standards

  Compatibility alias for engineering-guidelines. Show, check, or write coding-standards catalogs and optional module standards. Use when defining style/convention guidance or mapping conventions to lint and evaluate checks.

• ah-context

  Locate this project’s lifecycle and skills in central storage. Use when loading agenthouse guidance in any coding agent.

• ah-controls

  Show which policy rules map to real checks versus advisory text only. Use when asking what is enforced, how a rule is implemented, or which guidance has no automated control.

• ah-define-opportunity

  Describe an opportunity and its ideal customer profile: segments, buyer and user, job and pain, value hypothesis, channels, and pre-stated success and stop conditions. Use when starting market validation, defining an ICP, or before collecting customer evidence.

• ah-define-product-requirements

  Define product requirements and a story map from user evidence, assumptions, constraints, and missing information. Use when a request spans multiple stories or needs a PRD before story drafting.

• ah-demo

  Run a failing check, apply a fix, then a passing report in a new empty directory. Use when trying agenthouse for the first time or showing sample evaluation reports. Does not validate the user's application.

• ah-dependencies

  Show, pin, unpin, or update bundled skills such as frontend-acceptance and web-usability. Use when a specialist skill is missing, integrity fails, or applying a trusted dependency bundle.

• ah-doctor

  Diagnose broken enrollment, missing skills, policy snapshot problems, and housekeeping ignore gaps. Use when agenthouse looks unhealthy or skills fail to load.

• ah-draft-user-story

  Draft a story or bug with observable acceptance criteria, scope, and open questions. Use when writing requirements, turning an idea into a work item, or clarifying a bug.

• ah-engineering-guidelines

  Manage and consult engineering guidelines for architecture, testing, security, Git workflows, delivery, and coding conventions. Use when managing guidelines or assessing an approach against applicable guidance before acting.

• ah-enroll-repository

  Set up agenthouse in this repository by choosing coding agents, policy, and autonomy. Use when the project is not enrolled yet or when re-running conversational onboarding.

• ah-evaluate

  Run checks configured in the repository profile against a specific build and write evidence reports; does not assess requirements, scope or design. Use when running local or CI evaluation, pull-request checks, or verifying a specific build.

• ah-frontend-acceptance

  Compare UI work to the story, bug, wireframe, or reference image using real screenshots and the pinned frontend-acceptance method. Use when checking visual design, screens, or frontend acceptance.

• ah-gate

  Check ready or done criteria, evidence files, and independent signed approval for a work item. Use when asking if a story is ready to implement or done.

• ah-help

  List installed agenthouse skills and suggest which one fits a goal. Use when asking what agenthouse can do, which ah-* skill to pick, or for the command map.

• ah-hook-config

  Print, install, or remove agenthouse-owned hook entries in Claude/Cursor settings without touching unrelated hooks. Use when enabling or disabling agenthouse hooks in a supported host.

• ah-hook

  Run one coding-agent or CI hook event through the installed hooks runtime against frozen project policy. Use when processing a host hook payload, a CI event, or an explicit fixture file.

• ah-housekeep

  Verify that configured generated output stays out of Git; preserve fixtures, baselines and review evidence. Use when leftover captures appear, after frontend-acceptance, or at session start.

• ah-init

  Install the pinned runtime, lifecycle skill, and required frontend-acceptance dependency without the conversational questionnaire. Use when agents and autonomy are already chosen, or when reinstalling.

• ah-inspect

  List files touched by a commit or range, candidate tests, new suppressions, and leftover residue. Use when reviewing what a change touched or choosing which checks to run for a diff.

• ah-keygen

  Create an Ed25519 key pair at a requested path without overwriting existing files. Use when setting up signing keys for governance decisions or bundles.

• ah-lifecycle

  Guide a software change through the enrolled agenthouse lifecycle, linking its intended outcome, decisions, verification, release, operation, and learning. Use in repositories enrolled with agenthouse engineering when planning, implementing, reviewing, or delivering work.

• ah-market

  List, start, or check market-validation records (ICP segments, evidence, buyer signals, fit assessment) stored as Markdown in the repository. Use when starting an opportunity folder or validating that those files are complete before a go decision.

• ah-module

  Show a stack template, or preview/apply a reviewed plan that wires repository test commands into evaluate profiles. Use when connecting npm/php test scripts to agenthouse without hand-editing config.

• ah-npm-provenance

  Inspect local publish files for npm provenance, or write a GitHub Actions / GitLab job that enables it. Use when publishing to npm, setting up a trusted publisher, or asking whether a package should include provenance.

• ah-onboard

  Choose shared or private repository integration, coding agents, artifact locations, policy, and autonomy, then enroll this repository. Use when starting with agenthouse or adding it to a project.

• ah-pipeline

  Create and manage multi-job GitHub Actions or GitLab CI for evaluate profiles and publish test results. Use when wiring CI, adding frontend/release jobs, or fixing JUnit/artifact publishing.

• ah-policy

  Track signed organization policy origins, report drift and permission, or adopt reviewed revisions. Use when maintaining organization policy across repositories.

• ah-process

  Navigate the product lifecycle or subprocesses for features, bugs, changes, releases, and incidents, or request orientation about current evidence. Use when starting product-process guidance or asking where work stands. Process guidance does not track or change status.

• ah-record-market-evidence

  Record interviews, surveys, experiments, analytics, desk research, support feedback, and commercial signals as market evidence with sample and limitations. Use when adding customer or market research results to a market-validation record.

• ah-recover

  Finish an interrupted onboard, update, or restore without overwriting later user edits. Use when installation stopped mid-way or a lock remains.

• ah-resolve

  Rebuild or freeze-check the merged policy file at .agenthouse/resolved.json. Use when policy sources changed, verifying a frozen snapshot, or before evaluation that needs current rules.

• ah-restore

  Restore missing central assets or legacy project projections from this project’s exact recorded version. Use when the pinned runtime or skills are missing after a clone or machine change.

• ah-review-change

  Review a change against its requirements, policy, and available evidence, and report findings. Use when reviewing a pull request, diff, or implementation against a work item.

• ah-review-mr

  Review a PR or MR URL against its ticket in an isolated local environment and draft findings. Use when reviewing an incoming merge request without remote writes.

• ah-review

  Match each work-item acceptance criterion to build and policy evidence for an exact commit. Use when checking requirement coverage before merge; a technical pass is not independent review approval.

• ah-roles

  Discover and load role guidance, or adopt and review consumer-owned role baselines. Use when asking the agent to work from a role such as architect, product manager, engineer, test manager, devops, support, market research, or sales.

• ah-rollback

  Restore the previous complete runtime and dependency set. Use when an update should be undone and the prior version is still available.

• ah-session

  Report the active agenthouse version, apply approved updates and housekeeping, and flag policy changes without adopting them silently. Use at task start or when asking which version and policy are active.

• ah-sign

  Sign a reviewed governance decision or data bundle with an authorized private key. Use when a human has authorized signing; invoking this skill is not itself approval.

• ah-skill

  Import one reviewed specialist skill unchanged from a local folder. Use when adding an optional skill; required frontend-acceptance stays under dependencies.

• ah-spec

  Capture a failing (red) then passing (green) run of the same configured test command for a work item. Use when recording specification evidence that a check actually failed before the fix.

• ah-survey

  Read Git state, scripts, pipelines, and agent files to report which test layers look present, missing, or only nominally named—without running those commands. Use when exploring an unfamiliar repository or deciding what to wire into evaluate.

• ah-test-plan

  Propose and record which tests cover each acceptance criterion (level, method, file). Use when planning TDD or multi-level tests before writing them, including frontend interaction criteria.

• ah-uninstall

  Remove unchanged managed agenthouse files while keeping configuration, work records, and evidence. Use when removing the framework from a repository.

• ah-update

  Switch latest vs exact update tracking, or verify and install a reviewed agenthouse release (bundle or public channel). Use when changing update preference or applying an approved upgrade.

• ah-usability

  Download locked browser tooling if needed, then run the upstream web-usability audit against a URL or fixture. Use when collecting usability evidence; exit code alone is not a conformity certificate.

• ah-validate-scope

  Assess whether a set of requirements covers an intended outcome, including gaps, unrelated work, and sequencing. Use when checking if stories or requirement files are enough before implementation.

• ah-visual-plan

  Draft or validate wireframe HTML fragments and mermaid architecture diagrams for a work item. Use when planning UI layout, data models, or architecture before implementation, or checking an existing visual plan.

• ah-web-usability-conformity

  Run the installed web-usability audit and report real findings against a URL or fixture. Use when checking UI usability; this is separate from frontend-acceptance visual checks.

• ah-work

  Create, show, advance, or branch a local lifecycle work item under .agenthouse/work. Use when filing a story, opening a work record, moving it to a new stage, or creating a related Git branch.

• ah-write-tests

  Prepare to author failing tests from an accepted test plan after a green baseline, then mark criteria written and hand off to ah-spec red for TDD. Use when turning a test plan into repository tests without implementing product code.

Similar plugins