Code Repair Spine
by Corey Jacobs
Generate source-safe repository audits and repair handoff bundles without mutating your code.
- Generate risk-scored audit bundles before starting major refactors
- Create secret-redacted repo maps for safe LLM context injection
- Automate dependency graphing for Python and JS/TS projects
$19
· or 95 creditsSecure checkout via Stripe
Included in download
- Generate risk-scored audit bundles before starting major refactors
- Create secret-redacted repo maps for safe LLM context injection
- terminal automation included
- Ready for Codex
Sample input
Generate a repair spine audit for the current local directory and save the reports to ../audit_results.
Sample output
Code Repair Spine v0.1 Phase 1 complete.
Run folder: CRS_my-app_phase1_20231027-1430_a1b2c3d_MED-RISK
Bundle: 11 artifacts generated (SecretScan, DepGraph, BreakPlan, etc.)
Status: COMPLETED (read-only)
Next: Use _handoff/CodexBuildPrompt.md to begin repairs.
Code Repair Spine
by Corey Jacobs
Generate source-safe repository audits and repair handoff bundles without mutating your code.
$19
· or 95 creditsSecure checkout via Stripe
Included in download
- Generate risk-scored audit bundles before starting major refactors
- Create secret-redacted repo maps for safe LLM context injection
- terminal automation included
- Ready for Codex
- Instant install
Sample input
Generate a repair spine audit for the current local directory and save the reports to ../audit_results.
Sample output
Code Repair Spine v0.1 Phase 1 complete.
Run folder: CRS_my-app_phase1_20231027-1430_a1b2c3d_MED-RISK
Bundle: 11 artifacts generated (SecretScan, DepGraph, BreakPlan, etc.)
Status: COMPLETED (read-only)
Next: Use _handoff/CodexBuildPrompt.md to begin repairs.
About This Skill
High-Integrity Repository Auditing & Handoff
The Code Repair Spine (CRS) is a developer-centric audit tool designed to bridge the gap between discovery and execution. It generates a comprehensive, read-only Phase 1 bundle for local repositories, providing the evidence-backed planning needed before starting a repair or build pass. Unlike simple LLM prompts, this skill executes a multi-step workflow including dependency graphing, secret scanning, and risk assessment to build a structured context for any AI agent.
What it does
- Deep Inventory: Scans local repos to generate Python and JS/TS dependency graphs and repository maps.
- Secret Safety: Conducts secret scanning with mandatory redaction and salted fingerprints, so raw secret values don't enter your LLM context.
- Actionable Planning: Produces a "Break Plan," "Map of Change," and specific prompts (GPT/Codex/Claude/Hardening) tailored to the codebase.
- Read-Only Operation: Operates strictly in read-only mode against the target repo, and blocks any run whose output directory resolves inside that repo.
Why use this skill
Prompting an AI to "fix my code" often leads to hallucinations or missed architectural constraints. CRS provides a standardized "Spine"—a structured set of artifacts that ground the AI in reality. It automates the tedious work of repo mapping and risk analysis, delivering a handoff bundle that serves as a single source of truth for subsequent development phases.
Important boundary
This is not an auto-fixer, security certificate, compliance review, or production-readiness certificate. It does not claim that all secrets or all vulnerabilities were found, or that the repository is safe to deploy. It is a local, evidence-bound repo review and handoff tool for authorized repositories.
Use Cases
- Generate risk-scored audit bundles before starting major refactors
- Create secret-redacted repo maps for safe LLM context injection
- Automate dependency graphing for Python and JS/TS projects
- Produce evidence-backed 'Break Plans' for complex code migrations
- Auditing an inherited or unfamiliar repo before quoting or scoping repair work
- Prepping a grounded context bundle before a Claude/Codex/GPT repair session so the agent doesn't hallucinate architecture
- Generating a break plan + risk register before a refactor or migration
- Producing a reviewable handoff packet when passing a repo to another dev or agent
- Establishing an evidence baseline before a "fix my code" pass so the diff is bounded and reviewable
Known Limitations
Phase 1 only — not an auto-fixer. It plans and scopes the repair; it does not modify code or perform the fix itself. Static analysis only. Dependency graphs won't capture dynamic imports, runtime reflection, generated code, bundler aliases, or framework-specific resolution. Python and JS/TS dependency graphing only. Other languages get repo inventory and secret scanning, but no dependency graph. Secret scan is pattern-based. It flags likely exposures with redacted previews — it can miss novel or obfuscated secrets and may produce false positives. It is not a guarantee that all secrets were found. Not a security or compliance certificate. No claim of vulnerability-free, production-safe, or regulatory/compliance conformance. Generated prompts are scaffolding, not guaranteed-correct output. The follow-up GPT/Codex/Claude prompts ground the next pass; they don't guarantee the resulting code is correct. Does not replace human review. Warnings mean "look here," not "this is broken" or "this is safe."
How to Install
mkdir -p ~/.claude/skills && curl -sL https://www.agensi.io/api/install/code-repair-spine -o /tmp/code-repair-spine.zip && unzip -o /tmp/code-repair-spine.zip -d ~/.claude/skills && rm /tmp/code-repair-spine.zipFree skills install directly. Paid skills require purchase - use the download button above after buying.
Reviews
No reviews yet - be the first to share your experience.
Only users who have downloaded or purchased this skill can leave a review.
Early access skill
Be the first to review this skill.
Only users who have downloaded or purchased this skill can leave a review.
Security Scanned
Passed automated security review
Permissions
Allowed Hosts
File Scopes
Tags
Python 3.10+. Runs locally. No third-party runtime dependencies required. Best used with Claude Code, Codex, Cursor, VS Code, or any AI coding agent that can consume the generated handoff files.
Creator
Frequently Asked Questions
Learn More About AI Agent Skills
More Premium Skills

inline-comment
Best way to steer your agents, effortlessly.
Multi-Agent Orchestration Master Library
Transform Claude Code into a coordinated multi-agent system. Battle-tested tmux orchestration patterns, YAML task queues, event-driven communication, and parallel worker management for 8+ agents.

Legacy Code Modernization Planner for AI Coding Agents
Creates safe modernization roadmaps for old, messy, undocumented, or fragile codebases, including risk audits, refactor phases, dependency reviews, testing plans, migration steps, and AI coding prompts.
designing-hybrid-context-layers
Architects the right retrieval strategy for every query — teaching your agent when to use RAG, a knowledge graph, or a temporal index instead of defaulting to vector search for everything.