agentic-review-handoff
Agentic Review Handoff
Workflow
-
State the scope, then proceed.
- Name the exact scope: staged diff, working tree diff, full branch diff, generated artifacts, docs-only, or specific files.
- If scope is missing, reconstruct the minimum scope from the diff / file paths / prior findings the user gave you and proceed with explicit assumption labels. Asking a clarifying question is a last resort, only when no defensible scoped output is possible.
-
Use a handoff packet when the change is medium/large or the user asks for final review. See
references/handoff-packet.md. For tiny reviews, skip the packet but still state the scope. -
Review is read-only by default. Do not edit files, commit, push, or rebase unless the user explicitly switches to fix mode. Prefer code-level evidence over intent summaries. If verification is blocked, name what was not verified and why.
-
Output findings using the review contract in
references/review-contract.md— findings first, severity + file:line + Source tag + impact + fix per finding. If no issues, writeNo findingsplus checks run and residual risk. -
Handle review-fix-review loops.
- After fixes, review the changed fix scope and nearby regression surface.
- Do not restart a full review unless the user asks or the fix changes the architecture/scope.
More from adonis0123/adonis-skills
staged-changes-review
Checklist-driven review of staged Git changes with deterministic rule scanning and semantic analysis. Use when the user wants to review staged changes, check for errors before commit, or validate code quality before committing.
25ruler-rules-init
Migrate and bootstrap Ruler-based AI rules across repositories with English-first templates, preset-based project detection, and safe, idempotent setup. Use when creating or standardizing `.ruler/*` files, generating `AGENTS.md`/`CLAUDE.md`, wiring `ruler:apply` into `package.json`, adding generated-file ignores in `.gitignore`, and designing `applyTo` scoping patterns such as `**` and `web/**`.
18discuss-before-plan
Enforces 'decide then plan' discipline - the pre-planning decision gate. Use when the user asks for a plan or starts a change while key decisions are unresolved: architecture tradeoffs, data flow, public interfaces, unclear requirements, multi-module scope, or roughly 5+ files affected. Also triggers when the user explicitly wants to discuss, compare options, or review architecture before committing. Core job: reduce incorrect-execution cost by confirming decisions before producing executable plans.
17claude-skills-sync-init
Bootstrap Claude skill-sync automation in any repository. Use when users ask to set up `.agents/skills` to `.claude/skills` syncing, generate `skills:sync:*` scripts, inject package.json scripts/postinstall hooks, or initialize this sync workflow in a new or existing project.
16staged-review-validator
Validate staged-changes-review findings with evidence-based verdicts and minimal-fix guidance. Use when you need to confirm whether each reported issue is valid and actionable.
14tailwindcss-next-init
Initialize Tailwind CSS v4 and Iconify for Next.js App Router projects with script-driven templates. Use when you need a fast setup or migration with consistent style-entry conventions.
12