manifest-dev-tools

Utilities that sit alongside the define → do → verify workflow — prompt engineering, PR babysitting, PR walkthroughs and reviews, cross-boundary context handoff, and incremental teaching for session work.

Skills

Skill	Description
/babysit-pr	Author-side PR lifecycle babysitter and companion to /review-pr. Uses manifest grounding when available, synthesizes PR grounding when not, then runs the manifest lifecycle toward green and mergeable without pressing merge. Supports CI one-shot advancement via --ci; keeps a continuity journal by default under the user's home .manifest-dev/logs/ directory, with --no-log to disable it.
/handoff	Produce a self-contained context payload that lets a fresh agent continue without re-deriving understanding. Two triggers: cross-boundary transfer (tool switch, fresh session, another agent) and DIY sub-agent (spin off a focused side-session and hand back). Manually invoked.
/prompt-engineering	Create, update, or review an LLM prompt — system prompt, skill, or agent. State the goal, trust the model, add only what closes a real gap in natural behavior.
/review-pr	Autonomous PR review that posts high-signal, human-voiced comments under your account. Advances existing review threads, verifies fixes/replies/stale comments, and posts one GitHub review. Polymorphic on --manifest: without it, runs the generic reviewer fleet on the relevant diff range; with it, skips the fleet and independently verifies only the manifest — running each Acceptance Criterion and Global Invariant verify.prompt against the PR head (optionally cross-model) and posting PASS/FAIL. --loop schedules repeated one-shot passes with backoff.
/teach-me	Teach the learner to deeply understand a body of work — the current session, a PR, an ADR, or any topic. Builds a three-pillar checklist, teaches incrementally, and quizzes for demonstrated mastery before wrapping up.
/walk-pr	Walk through a PR or large diff together, one sub-changeset at a time.
review-prompt	Reviews LLM prompts against the /prompt-engineering skill's gap-calibration principles. Reports issues without modifying files, tagging each NEEDS_USER_INPUT or AUTO_FIXABLE so an optimization loop can act on them. Use when reviewing prompt quality, auditing a prompt, or evaluating a system prompt. A general-purpose verifier activates it from a verify.prompt.

How It Works

These tools sit alongside the manifest workflow (/define → /do → /done). /handoff produces a context payload for two use cases: cross-boundary transfer (tool switch, fresh session, multi-agent transfer) and DIY sub-agent flows (spin off a focused side-session and hand back to the parent without polluting its context). /teach-me turns a body of work — the session, a PR, an ADR, or any topic — into an incremental learning loop that verifies understanding before ending. /prompt-engineering, /walk-pr, /review-pr, and /babysit-pr are stand-alone collaboration tools — /walk-pr is the collaborative review surface, /review-pr is the autonomous reviewer, and /babysit-pr is the author-side PR lifecycle actor that orchestrates core manifest-dev skills.

Installation

/plugin install manifest-dev-tools@manifest-dev

For OpenCode, Codex, and Pi package installs, use the repo-level distribution instructions. Pi installs from the repository root and includes compatible shared tools skills plus /auto and /babysit-pr wrappers that route through the Pi Harness-level Do outcome gate.