- Home
- Skills
- Code Quality & Review
- Ouroboros Evidence Guardian
More screenshots
Works with the AI Agents you already use
Ouroboros Evidence Guardian
Audit Git changes, CI claims and deterministic-code risks with revision-bound evidence. Separate verified facts from assumptions and get actionable merge-readiness reports without fabricated runtime success.
$49
Ouroboros Evidence Guardian
Example session with this skill installed
Review this Git repository for deterministic simulation risks and tell me whether a merge is justified. Record the exact HEAD SHA, inspect tracked source files, separate verified findings from heuristic warnings, and explicitly list any CI or production checks you cannot verify. Do not modify the repository.
- Read your context and instructions
- Compiled the ouroboros evidence guardian
Summary
Decision: BLOCKED pending independent CI and runtime verification. The static scan completed against a fictional sample checkout and flagged two possible deterministic-state risks. A match does not prove either call influences game state.
Evidence scope
Repository: illustrative sample project
Revision: EXAMPLE_SHA (placeholder, not a real commit)
Worktree: assumed clean for this demonstration
Audit mode: read-only, local tracked-source heuristic
Severity
Surface
Location
Observation
Status
Medium
Core
src/simulation/tick.ts:14
Date.now() candidate; inspect whether wall-clock time reaches the authoritative tick
UNVERIFIED
Medium
Core
src/world/spawn.ts:31
Math.random() candidate; inspect whether a deterministic seed is required
UNVERIFIED
High
CI
No run receipt supplied
Exact-revision test results unavailable
BLOCKED
High
Runtime
No runtime receipt supplied
Live replay and tick invariants unverified
BLOCKED
Recommendations
Trace the Date.now() call site to its consumers. If it feeds authoritative simulation state, replace it with the existing deterministic tick abstraction.
Trace the random call. If it contributes to simulation output, use the system's reproducible seeded generator.
Run deterministic replay tests against the exact commit under review, record command, environment, exit code and output fingerprint.
Obtain an independent CI readback linked to that same SHA.
Reassess the merge only when required checks and real runtime boundaries are supported by evidence.
Limitations: This example is fictional and intentionally marked as such. No live repository, deployment, or CI was inspected for this listing demonstration.
Connects securely to your tools. The creator never sees your data.
About this skill
Stop treating a green badge as proof that software actually works. Ouroboros Evidence Guardian helps AI coding agents review Git repositories, pull requests, and release evidence with a consistent, revision-bound audit workflow. It records the exact Git SHA, inspects relevant source and checks, flags potential nondeterministic patterns, and separates verified observations from missing or blocked runtime proof. What you get A structured, reproducible read-only Git audit workflow. A bundled Python 3 standard-library scanner for tracked source files. Risk leads for Date.now(), Math.random() and related APIs. Evidence grading: VERIFIED / UNVERIFIED / FAILED / BLOCKED. Merge-readiness guidance that never treats static analysis alone as live runtime proof. A clear report format for findings, affected files, revision and next steps. Who it's for: AI-assisted developers, maintainers, technical leads, and teams using GitHub, GitLab or local Git repositories who want fewer unjustified "all good" claims. Important limitations: The included scanner is a heuristic, not a full static analyzer. CI, deployment and production evidence need access to the relevant systems. This product does not guarantee bug-free code, verify live production by itself, or automatically merge changes. No paid dependencies; Python 3 and Git required for the included scanner. Internet and account access are optional depending on which external evidence you ask the agent to verify.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Code Quality & Review.
Requires Python 3 for the scanner script. Cannot prove absolute correctness, only identifies heuristic leads. Audit is blocked if Git metadata or repository access is unavailable.
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
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 17 days ago
- Accounts
GitHub
Permissions
Allowed Hosts