human-alignment

Plugin human-alignment from marketplace agentic-engineering-patterns in memorysaver/agentic-engineering-patterns. Commands read from the repo's manifest — 11 days ago.

Skills (34)

• verify-aep

  Exercise an AEP preview binary through source-preserving migration and native inspection in a disposable project, retaining runtime evidence for release review.

• aep-design

  Designs a feature interactively on the integration branch into a reviewed OpenSpec proposal. Use for "design a feature" or "explore and propose".

• aep-git-ref

  Canonical AEP git reference: base resolution, worktree lifecycle, naming, task commits, PRs, recovery. Use when a git convention is needed.

• aep-launch

  Creates a designed feature's worktree and spawns its implementation agent, plus an optional evaluator. Use after /aep-design or for "launch workspace".

• aep-wrap

  Archives OpenSpec artifacts and removes the workspace after a PR merges. Use for "wrap up", "archive", or post-merge cleanup.

• aep-human-alignment

  Renders product-context.yaml into where we stand, what waits on a human, and what drifted. Use for pulse, status check, or a project brief; not planning.

• aep

  Use the AEP CLI to build project context, manage scoped work, and verify actual results in projects that use AEP. Read this when operating AEP or following a project's AEP entrypoint.

• bug-fix

  Fix a reported bug in AEP or a project: reproduce it, find the root cause, add a test that fails first, make the smallest fix, verify the committed candidate with the configured checks, and record a lesson when the same mistake repeats. Use for any bug whose fix changes code, tests or procedures.

• deliver

  Prepare a PR or perform authorized integration using current evidence, then preserve lessons and close owned resources.

• eval

  Watch another agent's work as a neutral observer (aep eval watch), judge AEP adherence and engineering practice from structural evidence, and record findings outside the project.

• implement

  Select ready work and implement an accepted story in its worktree, preserving task context, verification, and recovery information.

• migrate

  Convert useful legacy context into native stores, preserve source provenance, and explicitly switch the project's workflow owner.

• principles

  Read the admitted engineering principles before citing them in a runbook step or applying them to project work.

• project

  Install or inspect project entrypoints, assemble project context, and establish the project's runtime verification procedure and checks.

• reflect

  Turn observed product or execution feedback into context updates, lessons, project rules, local procedures, or grounded upstream proposals.

• roadmap

  Maintain product direction, journeys, collaboration containers, and decision records, or explain current progress from recorded evidence.

• stack

  Scaffold a new web project with Better-T-Stack and Vite+, native AEP specifications, local development, starter tests and configured checks.

• validate

  Verify implementation or other artifacts against intent using observable results, current evidence, and project-required checks or reviews.

• aep-autopilot

  Runs dispatch through wrap continuously for a whole layer, unattended. Use for "autopilot" or "hands-free"; for a single story use /aep-dispatch.

• aep-design-lens

  Reviews an existing UI against usability and accessibility heuristics, scored 0–4. Use for design or heuristic review; taste capture is /aep-calibrate.

• aep-easy-explain

  Stop. That last message did not land — re-pitch it.

• aep-executor

  Spawns and steers workspace agents across Claude Code and Codex backends. Use for backend choice, launch mode, or host detection; other skills call it.

• aep-gen-eval

  Defines the reusable generator/evaluator pattern: scoring, agent contracts, eval protocol. Validating one specific artifact is /aep-validate.

• aep-workflow-feedback

  Captures process learnings in a downstream project and routes them upstream to AEP. Use after /aep-wrap or /aep-reflect, or for "capture learnings".

• aep-workflow

  Authors a re-runnable multi-agent harness for one task too large or uncertain for a single context. Use for "dynamic workflow" or "orchestrate subagents".

• aep-calibrate

  Turns a human's subjective judgment into criteria agents can apply. Use for a works-vs-right gap, a .5 layer, or "calibrate" — not a usability audit.

• aep-dispatch

  Picks one ready story and hands it to /aep-design or /aep-launch. Use for "what's next" or "dispatch"; running continuously is /aep-autopilot.

• aep-envision

  Validates and frames a product opportunity into the Context Document downstream agents read. Use to start or revisit an idea, or for "what should we build".

• aep-map

  Decomposes a product into a system map, layered story graph, and agent topology. Use after /aep-envision, or for "break this down" and "plan the stories".

• aep-model

  Builds a human-reviewed OOUX object map and noun-first UI blueprint. Use after /aep-map, before UI stories, or for "object model" and "OOUX".

• aep-watch

  Ingests bug trackers, errors, and telemetry, then dedupes them into backlog stories. Use for "monitor for new work" or telemetry-driven stories.

• aep-e2e-skill-scaffolding

  Generates or upgrades a project's e2e-test skill to the three-tier BDD layer-gate shape. Use for "set up e2e testing" or "add BDD journeys".

• aep-onboard

  Installs, verifies, and explains AEP itself: environment, tools, plugin config. Use for first contact with AEP; creating a project is /aep-scaffold.

• aep-scaffold

  Creates a new AEP project or repairs an existing project's agentic infrastructure and skills layout. Installing AEP itself is /aep-onboard.