- Home
- Skills
- Agents & Orchestration
- contributor protocol
contributor protocol
Enforces a strict discovery and verification protocol for safe, documented, and approved codebase modifications.
Free
Works with the AI Agents you already use
contributor protocol
Example session with this skill installed
I need to rename the user internal ID field to uuid in the database schema and the API handler. What is the impact?
- Read your context and instructions
- Compiled the contributor protocol
I have analyzed the impact of renaming user_id to uuid. This change affects the core schema and the primary API handler.
Impact Analysis and Proposal
| Option | Risk | Effort | Impact | Recommendation |
|---|---|---|---|---|
| Rename field in schema and handler | Potential break for downstream consumers | M | Consistent naming convention | ✓ |
| Alias field in handler only | Technical debt in database schema | S | Lower risk of DB migration failure |
Proposed Changes:
src/schema/user.sql: Rename columnuser_idtouuid.src/api/handlers/user.ts: Update property reference fromuser_idtouuid.docs/api.md: Update field name in request/response examples.
Verification Plan:
- Run
npm run db:migrate:testto verify schema integrity. - Execute
npm testto ensure API handlers resolve the new property. - Check
CLAUDE.mdarchitecture section for stale references.
Proceed with these changes?
Next steps
- Approve the proposal to begin the multi-file edit.
- Provide the preferred migration script format if specific SQL dialects are required.
- Confirm if any external analytics services depend on the
user_idkey name.
Connects securely to your tools. The creator never sees your data.
About this skill
Standardizes the workflow for modifying codebases by enforcing a rigorous discover-propose-confirm loop. This skill eliminates silent failures and scope creep by requiring explicit approval for every file change and verifying results against the project's own build manifests and lockfiles. It treats project stability as the primary objective, ensuring that every modification rests on verified facts rather than model assumptions.
What it does
- Project Contract Extraction reads build manifests, lockfiles, and architecture docs to establish pinned versions and enforced invariants.
- Strict Confirmation requires a one-line proposal and explicit user approval before any file is edited.
- Impact Tracing analyzes dependency directions both ways to prevent cycles or regressions before proposing changes.
- Root Cause Enforcement prohibits bypassing safety checks, weakening tests, or suppressing linter warnings.
- Automated Definition of Done verifies changes against a checklist including test coverage, invariant validation, and documentation updates.
How it works
- Discovery phase where the agent reads
CLAUDE.md,package.json, or similar manifests to understand the environment. - Proposal phase where the agent presents a specific change set and asks for approval.
- Application phase where only the approved diffs are applied to the codebase.
- Verification phase where the agent runs project-specific test suites and reports actual results.
Frameworks & tools
Works with any language or framework including React, Node.js, Python, Go, and Rust. It uses project files like CLAUDE.md, README.md, and standard build tool outputs to drive decisions.
Why this beats prompting it yourself
Generic prompts often lead to "hallucinated" API usage or silent regressions in complex dependencies. This skill forces the agent to verify every external API against pinned versions and prevents the rewriting of Git history or the introduction of architectural debt.
Use cases
- Safely refactoring legacy code while maintaining enforced invariants.
- Implementing features in a team environment where
CLAUDE.mdrules must be strictly followed. - Fixing bugs without introducing new linter warnings or breaking public API contracts.
Known limitations
Does not support rewriting Git history (rebase/amend). Verification is limited to commands available in the local environment and cannot verify UI behavior or hardware-specific integrations.
Prohibits rewriting Git history. Cannot verify behavior requiring external credentials, network access, or specialized hardware not present in the local runtime.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 18 days ago
- Free with an Agensi account
Requires