pr-review-run
Run a comprehensive multi-agent PR review on code changes. Use when the user asks to review a PR, review code changes, review a branch, run a code review, or check code before merging. Supports git diff comparisons, specific file review, and directory-level review. Orchestrates 7 specialized agents (architecture, code quality, bugs, security, tests, performance, acceptance) in parallel and produces an aggregated report with verdict.
What this skill does
# PR Review Orchestrator
## Workflow
### Step 1: Gather Review Context
Collect the review target from `$ARGUMENTS`.
**Supported input modes:**
1. **Git diff** — branch comparison (e.g. `main..HEAD`)
2. **Specific files** — list of file paths
3. **Directory** — all source files in a directory
If no arguments provided, use `AskUserQuestion` to ask:
- What branch or files to review
- Whether there is a task/ticket with requirements (needed for Acceptance Checker agent)
### Step 2: Collect Code Changes
**Git diff mode:**
```bash
git diff <base-branch>..HEAD --name-only # changed file list
git diff <base-branch>..HEAD # full diff
```
**Specific files:** Read each file directly.
**Directory:** List and read source files, excluding `node_modules`, `dist`, `build`, `.git`, etc.
### Step 3: Select Agents
**Core agents (always run):**
- `pr-review:code-cleaner`
- `pr-review:bug-smasher`
**Conditional agents:**
| Agent | Trigger Condition |
|-------|-------------------|
| `pr-review:architect-visioner` | New files > 2 OR lines changed > 100 OR new module/class |
| `pr-review:acceptance-checker` | Task requirements provided by user |
| `pr-review:security-guard` | Files contain: auth, login, password, token, api, input, form, sql, query, env, secret, key |
| `pr-review:test-guardian` | Test files in diff (*.test.*, *.spec.*, __tests__/*) |
| `pr-review:performance-scout` | Files contain: database, query, loop, map, filter, reduce, fetch, api, cache |
After auto-selection, show the user which agents are selected and which are skipped (with reasons). Use `AskUserQuestion` with `multiSelect` to let the user override the selection.
### Step 4: Run Agents in Parallel
**CRITICAL:** Run ALL selected agents in a SINGLE message using multiple Task tool calls.
Each agent is already registered as a sub-agent (e.g. `pr-review:code-cleaner`). Use the agent name directly as the `subagent_type` in the Task tool — their definitions are loaded automatically as the system prompt. Do NOT manually search for or read agent definition files.
Provide each agent with this prompt:
```
Review the following code changes:
## Files Changed
[list of files]
## Code Diff
[full diff or file contents]
## Project Context
[CLAUDE.md content if it exists]
```
For `pr-review:acceptance-checker`, also include the task requirements from step 1.
### Step 5: Aggregate Results
1. Collect all findings from completed agents
2. Deduplicate overlapping issues (apply deduplication rules from [references/output-format.md](references/output-format.md))
3. Sort by severity: Critical → High → Medium → Low
4. Calculate totals for the summary table
5. Determine verdict using the decision matrix from [references/output-format.md](references/output-format.md)
### Step 6: Format Report
Format the aggregated findings using the report template and rules from [references/output-format.md](references/output-format.md). See [references/example-report.md](references/example-report.md) for a complete example.
### Step 7: Save Report
Save the formatted report to:
```
docs/pr-reviews/{branch-name}-{YYYY-MM-DD}.md
```
Create the `docs/pr-reviews/` directory if it does not exist.
### Step 8: Present Results and Offer Actions
Display the executive summary (verdict + severity counts + report path).
Use `AskUserQuestion` to offer:
1. **View full report** — display the complete report
2. **Apply suggested fixes** — for Critical/High issues
3. **Done** — no further action
If "Apply suggested fixes" is selected:
- Show each Critical/High issue with its suggested fix
- Let the user pick which fixes to apply
- Apply selected fixes to the code
## Error Handling
- If an agent fails, continue with the remaining agents and note the failure in the report
- If `git diff` fails, fall back to asking the user for specific file paths
- If no code changes are found, inform the user and exit gracefully
Related in AI Agents
skill-development
IncludedComprehensive meta-skill for creating, managing, validating, auditing, and distributing Claude Code skills and slash commands (unified in v2.1.3+). Provides skill templates, creation workflows, validation patterns, audit checklists, naming conventions, YAML frontmatter guidance, progressive disclosure examples, and best practices lookup. Use when creating new skills, validating existing skills, auditing skill quality, understanding skill architecture, needing skill templates, learning about YAML frontmatter requirements, progressive disclosure patterns, tool restrictions (allowed-tools), skill composition, skill naming conventions, troubleshooting skill activation issues, creating custom slash commands, configuring command frontmatter, using command arguments ($ARGUMENTS, $1, $2), bash execution in commands, file references in commands, command namespacing, plugin commands, MCP slash commands, Skill tool configuration, or deciding between skills vs slash commands. Delegates to docs-management skill for official documentation.
reprompter
IncludedTransform messy prompts into well-structured, effective prompts — single or multi-agent. Use when: "reprompt", "reprompt this", "clean up this prompt", "structure my prompt", rough text needing XML tags and best practices, "reprompter teams", "repromptception", "run with quality", "smart run", "smart agents", multi-agent tasks, audits, parallel work, anything going to agent teams. Don't use when: simple Q&A, pure chat, immediate execution-only tasks. See "Don't Use When" section for details. Outputs: Structured XML/Markdown prompt, quality score (before/after), optional team brief + per-agent sub-prompts, agent team output files. Success criteria: Single mode quality score ≥ 7/10; Repromptception per-agent prompt quality score 8+/10; all required sections present, actionable and specific.
adaptive-compaction
IncludedAdaptive add-on policy and recovery layer that decides WHEN to compact, prune, snapshot, or fork -- replacing fixed-percent auto-compaction across Claude Code, Codex, and MCP-capable hosts. Trigger on auto-compact timing or damage: "when should I compact", "is it safe to compact now or start a fresh session", "auto-compact fires too early/mid-task", "switching to an unrelated task but the window still has space", "context rot", "answers get worse the longer the session runs", "the agent forgot the plan or my decisions after it summarized", "add a layer on top that manages context without changing the agent", raising autoCompactWindow to give the policy room, or installing/tuning a cross-tool compaction policy or PreCompact hook -- even when "compaction" is never said but the problem is context-window pressure or post-summarization memory loss. Do NOT use to summarize a conversation, build RAG, write a summarization prompt (decides WHEN not HOW), or answer max-context-length trivia.
agent-skill-creator
IncludedCreate cross-platform agent skills from workflow descriptions. Activates when users ask to create an agent, automate a repetitive workflow, create a custom skill, or need advanced agent creation. Triggers on phrases like create agent for, automate workflow, create skill for, every day I have to, daily I need to, turn process into agent, need to automate, create a cross-platform skill, validate this skill, export this skill, migrate this skill. Supports single skills, multi-agent suites, transcript processing, template-based creation, interactive configuration, cross-platform export, and spec validation.
llm-wiki
IncludedUse when building or maintaining a persistent personal knowledge base (second brain) in Obsidian where an LLM incrementally ingests sources, updates entity/concept pages, maintains cross-references, and keeps a synthesis current. Triggers include "second brain", "Obsidian wiki", "personal knowledge management", "ingest this paper/article/book", "build a research wiki", "compound knowledge", "Memex", or whenever the user wants knowledge to accumulate across sessions instead of being re-derived by RAG on every query.
skill-master
IncludedAgent Skills authoring, evaluation, and optimization. Create, edit, validate, benchmark, and improve skills following the agentskills.io specification. Use when designing SKILL.md files, structuring skill folders (references, scripts, assets), ingesting external documentation into skills, running trigger evals, benchmarking skill quality, optimizing descriptions, or performing blind A/B comparisons. Keywords: agentskills.io, SKILL.md, skill authoring, eval, benchmark, trigger optimization.