Automated summary
Starts a technical review session based on an existing plan and implementation, by discovering available plans and invoking the review skill.
Strengths
- Automatically discovers available plans
- Ensures traceability between plan and implementation
- Part of a structured six-phase workflow
- Runs necessary migrations before review
Limitations
- Requires plans and implementation to already exist
- Does not perform the review itself, only sets it up
- Relies on a discovery script for current state
Use this skill when you have completed an implementation based on a plan and want a structured review.
Do not use it if no plan exists or if you need an informal review without reference to a plan.
Security analysis
SafeThe skill runs a fixed bash script for discovery and then invokes another skill; there are no destructive, exfiltrating, or obfuscated actions. It does not execute arbitrary user input.
No concerns found
How this analysis is madeExamples
Start a review session for the current implementation based on the existing plan.Let's review the implemented code against the plan we created earlier.Invoke the technical review skill to validate the implementation.name: start-review description: "Start a review session from an existing plan and implementation. Discovers available plans, validates implementation exists, and invokes the technical-review skill." disable-model-invocation: true allowed-tools: Bash(.claude/scripts/discovery-for-implementation-and-review.sh)
Invoke the technical-review skill for this conversation.
Workflow Context
This is Phase 6 of the six-phase workflow:
| Phase | Focus | You | |-------|-------|-----| | 1. Research | EXPLORE - ideas, feasibility, market, business | | | 2. Discussion | WHAT and WHY - decisions, architecture, edge cases | | | 3. Specification | REFINE - validate into standalone spec | | | 4. Planning | HOW - phases, tasks, acceptance criteria | | | 5. Implementation | DOING - tests first, then code | | | 6. Review | VALIDATING - check work against artifacts | ◀ HERE |
Stay in your lane: Verify that every plan task was implemented, tested adequately, and meets quality standards. Don't fix code - identify problems. You're reviewing, not building.
Instructions
Follow these steps EXACTLY as written. Do not skip steps or combine them. Present output using the EXACT format shown in examples - do not simplify or alter the formatting.
CRITICAL: This guidance is mandatory.
- After each user interaction, STOP and wait for their response before proceeding
- Never assume or anticipate user choices
- Even if the user's initial prompt seems to answer a question, still confirm with them at the appropriate step
- Complete each step fully before moving to the next
- Do not act on gathered information until the skill is loaded - it contains the instructions for how to proceed
Step 0: Run Migrations
This step is mandatory. You must complete it before proceeding.
Invoke the /migrate skill and assess its output.
If files were updated: STOP and wait for the user to review the changes (e.g., via git diff) and confirm before proceeding to Step 1. Do not continue automatically.
If no updates needed: Proceed to Step 1.
Step 1: Run Discovery Script
Run the discovery script to gather current state:
.claude/scripts/discovery-for-implementation-and-review.sh
This outputs structured YAML. Parse it to understand:
From plans section:
exists- whether any plans existfiles- list of plans with: name, topic, status, date, format, specification, specification_exists, plan_id (if present)count- total number of plans
From state section:
scenario- one of:"no_plans","single_plan","multiple_plans"
IMPORTANT: Use ONLY this script for discovery. Do NOT run additional bash commands (ls, head, cat, etc.) to gather state - the script provides everything needed.
→ Proceed to Step 2.
Step 2: Route Based on Scenario
Use state.scenario from the discovery output to determine the path:
If scenario is "no_plans"
No plans exist yet.
No plans found in docs/workflow/planning/
The review phase requires a completed implementation based on a plan. Please run /start-planning first to create a plan, then /start-implementation to build it.
STOP. Wait for user to acknowledge before ending.
If scenario is "single_plan" or "multiple_plans"
Plans exist.
→ Proceed to Step 3 to present options.
Step 3: Present Plans and Select
Present all discovered plans to help the user make an informed choice.
Present the full state:
Available Plans:
1. {topic-1} ({status}) - format: {format}, spec: {exists|missing}
2. {topic-2} ({status}) - format: {format}, spec: {exists|missing}
· · ·
Which plan would you like to review the implementation for? (Enter a number or name)
If single plan exists (auto-select):
Auto-selecting: {topic} (only available plan)
→ Proceed directly to Step 4.
If multiple plans exist:
STOP. Wait for user response.
→ Based on user choice, proceed to Step 4.
Step 4: Identify Implementation Scope
Ask the user what code to review:
· · ·
What code should I review?
1. All changes since the plan was created
2. Specific directories or files
3. Let me identify from git status
Which approach?
STOP. Wait for user response.
If they choose specific directories/files, ask them to specify.
→ Proceed to Step 5.
Step 5: Invoke the Skill
After completing the steps above, this skill's purpose is fulfilled.
Invoke the technical-review skill for your next instructions. Do not act on the gathered information until the skill is loaded - it contains the instructions for how to proceed.
Example handoff:
Review session for: {topic}
Plan: docs/workflow/planning/{topic}.md
Format: {format}
Plan ID: {plan_id} (if applicable)
Specification: {specification} (exists: {true|false})
Scope: {all changes | specific paths | from git status}
Invoke the technical-review skill.
From the file to a first result
The download contains exactly the content displayed on this page. This copy may differ from the latest version in the source repository.
View the source and supporting filesThis download contains SKILL.md only. Any scripts, references, templates and dependencies are not included. Get the required files from the source, preserving their directory structure, before continuing.
1. Download the file
Download SKILL.mdFrom the root of a test project, save the file as .skill-downloads/start-a-review-session-mmlj2ouo/SKILL.md, or use the command below. This folder is for inspecting the skill before installation.
Command for macOS/Linux or WSL: requires a POSIX shell and curl. On Windows without WSL, use the download button and create the folder manually.
mkdir -p '.skill-downloads/start-a-review-session-mmlj2ouo' && curl --fail --show-error --location 'https://skills-guides.com/api/skills/start-a-review-session-mmlj2ouo/download' --output '.skill-downloads/start-a-review-session-mmlj2ouo/SKILL.md.part' && mv '.skill-downloads/start-a-review-session-mmlj2ouo/SKILL.md.part' '.skill-downloads/start-a-review-session-mmlj2ouo/SKILL.md'2. Check prerequisites
Read SKILL.md before using it. If it references scripts or external services, inspect them and prepare the required access in your test environment.
Tools declared by the author (check availability in your agent): Bash(.claude/scripts/discovery-for-implementation-and-review.sh)
No recognized prerequisites section in this file. This does not mean it is self-contained: libraries, tools and access requirements may be mentioned elsewhere in the content or source repository.
The platforms on this page are catalog indications. The paths below follow tool documentation; they do not prove this skill has been run in those tools. No minimum tool version is confirmed for this listing.
3. Load the skill
After inspection, copy the complete folder into your project using one of the commands below, then open a new tool session. The command stops if a folder with the same name already exists: compare the files before replacing it.
Claude Code
test -f '.skill-downloads/start-a-review-session-mmlj2ouo/SKILL.md' && test -s '.skill-downloads/start-a-review-session-mmlj2ouo/SKILL.md' && mkdir -p '.claude/skills' && test ! -e '.claude/skills/start-review' && cp -R '.skill-downloads/start-a-review-session-mmlj2ouo' '.claude/skills/start-review'OpenSkills
OpenSkills must already be installed. The local install/read procedure is checked with OpenSkills 1.5.0, which requires Node.js ≥ 20.6.0. Catalog slugs are not resolved automatically.
Official OpenSkills installationopenskills install './.skill-downloads/start-a-review-session-mmlj2ouo'openskills read 'start-a-review-session-mmlj2ouo'Reading should display the installed file content and its base directory. It does not run the skill or install its dependencies.
For an agent that reads AGENTS.md, then sync the index. This command changes AGENTS.md: review the diff before using it.
openskills syncManual method or another tool: attach SKILL.md to a conversation, or open it in an agent that can read files. Include any required supporting resources. No universal folder is recognized by every tool.
4. Try a first task
Use a short, synthetic input: an API endpoint, a small document or an interface brief. In a new conversation, attach the file or give the agent its path, then adapt this request:
Read .skill-downloads/start-a-review-session-mmlj2ouo/SKILL.md and apply its instructions to this example: [describe a test input]. Identify any missing files, dependencies or access. Produce the result requested by the skill without inventing data or running unreviewed scripts.Expected result: output matching the skill instructions, or a precise list of missing prerequisites. A read confirmation alone does not validate a first use. If an external resource is unavailable, that path remains untested.
ClickHouse Performance Comparison
Testing
Evaluate ClickHouse performance test results, check PR performance changes, compare with history, and assess whether a result is actionable.
Analyze CI Performance Reports
Testing
Analyze CI performance comparison reports for a ClickHouse PR. Cross-references master history to distinguish real regressions from flaky tests.
Santa Method
Testing
Multi-agent adversarial verification with two independent reviewers. Output is checked twice and iteratively fixed until both pass before shipping.