- Home
- Skills
- Agents & Orchestration
- Rita planning framework
Rita planning framework
Orchestrate feature planning, quality scoring, and doc-to-code drift auditing within the Rita framework.
Free
Works with the AI Agents you already use
Rita planning framework
Example session with this skill installed
Audit the features/billing-engine folder against the code in the billing/ directory to find any drift or undocumented changes.
- Read your context and instructions
- Compiled the rita planning framework
I have performed a delta audit of the features/billing-engine folder. The mechanical pass flagged two stale files based on recent commits to billing/processor.py, and the semantic pass identified a mismatch in the rate-limit implementation.
┌─ Rita delta report ─ features/billing-engine [mode: post-ship]
│ status Shipped · last reviewed 2024-11-12 (92d ago)
│ mechanical 2 docs STALE (drift_check.py)
│ consistency 1 doc↔code mismatch, 1 undocumented behaviour
│ open items 0 unresolved :warning:
│ hygiene 1 stale review date
└─
Findings
1. Consistency: Rate Limit Mismatch
- Doc claim:
README.md:45states a maximum burst of 500 requests. - Code reality:
billing/limiter.go:12definesMaxBurst = 250. - Fix options: Update
README.mdto reflect 250, or adjustlimiter.goto 500.
2. Undocumented Behavior: Retry Logic
- Code reality:
billing/client.go:88implements exponential backoff for 503 errors. - Doc gap: No mention of retry strategies in
README.mdorrunbook.md. - Fix options: Add the retry policy to the "System Interactions" section of the README.
3. Mechanical: Stale Review Date
metrics.mdis marked as last reviewed on 2024-11-12, butbilling/telemetry.gohas 4 commits since that date affecting metric names.
Next steps
- Choose whether to update the documentation or the implementation for the rate limit mismatch.
- Run
rita-planto document the existing retry logic in the feature README. - Update the
Last reviewed:date in the frontmatter after verifying the metrics implementation.
Connects securely to your tools. The creator never sees your data.
About this skill
Rita (Release It Toolkit) is a specialized agent package for managing the lifecycle of non-trivial features, from initial planning to long-term maintenance. It solves the problem of documentation rotting as soon as code is written and ensures that architectural intent remains visible throughout the implementation process. Developers use Rita to bootstrap feature folders, conduct feasibility checks, and audit documentation against implementation to prevent mechanical and semantic drift.
What it does
- Feature Bootstrapping creates structured feature folders with templates for READMEs, feasibility studies, metrics, and runbooks.
- Drift Detection audits existing documentation against the codebase to identify stale claims, missing metrics, or undocumented behaviors.
- Automated Plan Review runs a rubric-based scorecard and checklist to verify plan quality and framework compliance before human review.
- Feasibility Verification documents load-bearing assumptions with observed exit codes and outputs to prevent planning on false premises.
- Status Hygiene identifies stalled features by comparing doc frontmatter against actual commit history and implementation progress.
How it works
- rita-plan initiates the lifecycle by bootstrapping a feature folder and walking through the six phases of planning and shipping.
- rita-review performs an internal consistency check of the docs against a standard rubric, providing a quality scorecard.
- rita-delta runs
drift_check.pyand semantic analysis to find doc-to-code gaps and proposes specific fix options.
Frameworks & tools
This skill works with any codebase stored in a Git repository. It uses Python for the drift_check.py mechanical audit and relies on standard shell commands for file management.
Why this beats prompting it yourself
Generic prompts often hallucinate feasibility results or fail to catch subtle inconsistencies between a README and a Runbook. Rita enforces a specific lifecycle and uses subagents for fresh-context reviews, ensuring the agent evaluates what is actually on the page rather than its own authoring intent.
Use cases
- Drafting a new system feature while ensuring all invariants and constraints are documented.
- Auditing a legacy feature folder to see if the current code still matches the original metrics and runbook.
- Performing a pre-ship audit to confirm the implementation matches the defined Definition of Done.
Known limitations
Mechanical drift detection requires a Git repository for commit-based staleness checks. It cannot autonomously perform code refactors or flip feature flags without human approval.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
Requires Git for mechanical drift checks. Read-only auditing; does not automatically mutate code or docs. Detection of identifiers built at runtime may require human adjudication.
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 buying a skill.
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 to use
No account required to browse
Trust & safety
Security scan
Security scan passed 8 days ago
- Free with an Agensi account
- Software
Git