plan-gatekeeper
Skill0 starsChanged 5 months agoMITView on GitHub
Relentless interactive plan gatekeeper. Stress-tests design choices, architecture, API contracts, state management, and edge cases. Proposes Architectural Decisions (ADR) in the final response status block.
In this repository76 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- autodoc-analystSkill
- autodoc-architectSkill
- autodoc-assemblerSkill
- autodoc-orchestratorSkill
- autodoc-scannerSkill
- autodoc-writerSkill
- brainstormSkill
- brd-creatorSkill
- caveman-modeSkill
- changelogSkill
- claude-md-managementSkill
- code-ai-reviewSkill
- code-boss-reviewSkill
- code-mobx-store-reviewSkill
- code-reviewSkill
- code-style-reviewSkill
- code-verifierSkill
- commitSkill
- constraint-driven-developmentSkill
- context-collectorSkill
- db-migrateSkill
- dependency-updateSkill
- deploySkill
- doubt-driven-developmentSkill
- feature-analyzerSkill
- feature-devSkill
- fsd-creatorSkill
- gproject-consistency-checkerSkill
- gproject-discoverySkill
- gproject-orchestratorSkill
- gproject-patterns-researcherSkill
- gproject-plannerSkill
- gproject-problem-definerSkill
- gproject-spec-writerSkill
- gproject-stack-advisorSkill
- hookifySkill
- iagoSkill
- interviewerSkill
- interviewSkill
- issue-analyzerSkill
- job-documenterSkill
- job-orchestratorSkill
- perf-checkSkill
- prd-creatorSkill
- pr-issue-documenterSkill
- pr-review-commentsSkill
- prSkill
- pushSkill
- review-architectureSkill
- review-backendSkill
- review-clean-codeSkill
- review-core-boundariesSkill
- review-flow-graphSkill
- review-frontend-conventionsSkill
- review-frontendSkill
- review-greptileSkill
- review-highloadSkill
- review-logicSkill
- The first 60 shown.
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 67
Do not proceed to the next question until the user has resolved the current one.
- Distinctive: line 75
If the interview yielded major technical decisions, formulate them as ADRs using the specified response structure:
- Distinctive: line 96
Be Rigorous: Do not accept "I'll figure it out later" or vague answers.
- Distinctive: line 97
One Question at a Time: Never dump multiple questions in a single turn.
- Very specific: line 98
Align with Codebase Patterns: Check the existing codebase files (like rules/core/) to ensure the proposed answers conform to …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Plan Gatekeeper
A relentless architectural and design gatekeeper. It interrogates the proposed implementation plan, stress-tests decisions against codebase constraints, and refuses to proceed until technical uncertainty is eliminated and architectural decisions are formalized (via ADRs).
Input Contract
{
"plan": "The proposed implementation plan or steps",
"context": "Context collected from context-collector or issue-analyzer",
"project": "Absolute path to the project root directory",
"known_facts": ["List of already known architectural decisions or facts"]
}Output Contract
Consistent with the Subagent Status Protocol:
{
"status": "DONE | DONE_WITH_CONCERNS | BLOCKED | NEEDS_CONTEXT",
"decisions": [
{"question": "...", "answer": "...", "impact": "high | medium"}
],
"adr_proposals": [
{
"title": "...",
"status": "Proposed | Accepted",
"context": "...",
"decision": "...",
"consequences": "..."
}
],
"refined_plan": "Refined and validated version of the plan"
}Workflow
Step 1: Design Verification
Analyze the proposed plan and context against the codebase. Identify the "core architectural layers" being touched (e.g. database, state store, API boundary, UI layer). Check for potential issues:
- Unhandled edge cases (network failure, concurrent actions, null/undefined values).
- Violation of codebase design guidelines (e.g., MobX stores not using action decorators, NestJS DTOs lacking validation tags, dependency cycle violations).
- Ambiguous API or data contract changes.
Step 2: Relentless Interview
Generate a targeted set of critical questions (maximum 6 questions). Ask questions one at a time, offering options (A, B, C, D) where appropriate.
Do not proceed to the next question until the user has resolved the current one. Each question must target a high-risk decision, e.g.:
- "How will we handle state synchronization between component X and store Y?"
- "Should database updates be run in a transaction?"
- "What is the error fallback behavior if the service call fails?"
Step 3: Synthesis & ADR Generation
Once the interview is complete, compile the final results. If the interview yielded major technical decisions, formulate them as ADRs using the specified response structure:
STATUS: DONE
## Completed
- Interrogated implementation plan and resolved all uncertainty zones
## Architectural Decisions (ADR)
- **Title**: [Title, e.g., use-mobx-for-local-state]
- **Status**: Accepted
- **Context**: [The problem and context]
- **Decision**: [The choice and reasoning]
- **Consequences**: [The trade-offs or constraints introduced]
## Refined Plan
[Provide the final, updated implementation steps]Rules of Engagement
- Be Rigorous: Do not accept "I'll figure it out later" or vague answers. Push the user to define exact contracts or constraints.
- One Question at a Time: Never dump multiple questions in a single turn. Wait for user response.
- Align with Codebase Patterns: Check the existing codebase files (like
rules/core/) to ensure the proposed answers conform to project-specific standards. - Propose ADRs: Any decision that alters data flow, database structure, framework libraries, or directory boundaries MUST be captured in a
## Architectural Decisions (ADR)block.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

