- Home
- Skills
- Code Quality & Review
- Code Path Security Review
Works with the AI Agents you already use
Code Path Security Review
Trace untrusted data from entry point to sink with evidence, not vibes
$18
Code Path Security Review
Example session with this skill installed
Review this authentication callback diff for security issues. Focus on untrusted input flow, reachable sinks, and missing enforcement. Do not modify the code. Report only findings supported by the supplied implementation.
- Read your context and instructions
- Compiled the code path security
Illustrative synthetic demo. It demonstrates the expected review structure. No production repository, external target, or real vulnerability is implied.
Secure Code Review Record
SCOPE
Synthetic authentication callback diff.
REVIEW MODE
Read only source analysis.
Finding SCR-01
Class
Open redirect caused by untrusted redirect input.
Entry point
A synthetic next query parameter is accepted from the request.
Data flow
The worked example models the value moving from the request parameter into the post authentication redirect path without an allowlist or same origin validation step.
Sink
Application redirect function.
Control observed
No destination validation is present in the synthetic code path.
Reachability
SUPPORTED IN SYNTHETIC EXAMPLE
The modeled callback invokes the redirect with the user controlled value.
Impact
An attacker could construct a login link that sends the user to an attacker controlled destination after authentication.
Confidence
High within the illustrative code path.
Recommended control
Resolve the destination against an explicit internal route policy or approved same origin allowlist before redirecting.
Reviewed But Not Confirmed
No SQL execution, filesystem operation, server side URL fetch, or unsafe deserialization sink is present in the supplied synthetic review surface.
These vulnerability classes are therefore not reported as findings.
Review Boundary
No source files were modified.
No runtime exploit was attempted.
No claim is made about files or execution paths outside the supplied synthetic scope.
Next Step
Implement the approved redirect validation at the enforcement boundary, then verify the corrected behavior with a focused security test.
Connects securely to your tools. The creator never sees your data.
About this skill
Performs a security focused, read only review of a code diff or bounded implementation surface. It traces untrusted data from entry points to security sensitive sinks, checks whether relevant controls actually intervene, and reports only findings supported by reachable code paths and concrete evidence.
What it does
- Reviews a defined diff, pull request, module, or bounded code surface.
- Traces untrusted input from source to sink before classifying a vulnerability.
- Inspects reachable injection, XSS, SSRF, path traversal, unsafe deserialization, mass assignment, logic bypass, and fail open patterns.
- Evaluates validation, authorization, encoding, sanitization, and boundary enforcement along the actual execution path.
- Distinguishes exploitable code paths from dead code, unreachable branches, and unsupported scanner signals.
- Produces evidence backed findings with affected paths, causal mechanism, impact, and remediation direction.
- Remains read only and does not patch the reviewed implementation.
How it works
- Scope definition identifies the exact diff, files, module, or implementation surface to review.
- Entry point analysis identifies attacker controlled or otherwise untrusted inputs.
- Data flow tracing follows relevant values through transformations, validation, and trust boundaries.
- Sink analysis determines whether a security sensitive operation is reachable.
- Control analysis checks whether the relevant protection exists at the correct boundary.
- Finding classification records the causal path, evidence, confidence, impact, and recommended next workflow.
Boundaries
This skill reviews application code for security weaknesses.
Authentication architecture, sessions, and access models that require a dedicated identity review belong to identity-access-review. Secrets, sensitive data handling, and cryptographic design belong to data-secrets-crypto-review. Runtime and deployment configuration belongs to security-config-review.
A finding that requires active exploitability proof belongs to exploit-validation. Implementation of an approved fix belongs to security-fix-implementer.
Why this beats prompting it yourself
Generic code review prompts often flag dangerous function names without proving that attacker controlled data can actually reach them. They also tend to mix code review with configuration, identity design, and active exploitation. This skill keeps the review bounded, traces real data flow, checks the actual enforcement boundary, and reports findings only when the code path supports them.
Validation
The skill has been applied in independent read only review workflows and is also covered by JetFerrari Security Engineering pack level structural, routing, installation, discovery, and behavioral validation.
Use cases
- Review a pull request for security relevant regressions.
- Trace user controlled input to a sensitive sink.
- Investigate suspected injection, SSRF, path traversal, or unsafe deserialization.
- Separate reachable security defects from static analysis noise.
- Produce evidence backed findings before remediation planning.
Known limitations
This is a read only code review workflow. It does not modify source files, deploy changes, or actively exploit a target. Findings depend on the code and project context available to the agent. Runtime behavior that cannot be established from the reviewed evidence remains unverified and may require focused security testing or exploit validation.
Part of JetFerrari Security Engineering.
This is a read only code review workflow. It does not modify source files, deploy changes, or actively exploit a target. Findings depend on the code and project context available to the agent. Runtime behavior that cannot be established from the reviewed evidence remains UNVERIFIED and may require focused security testing or exploit validation.
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
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 24 days ago
Permissions
Access is limited to the code, diff, and project context supplied for the review. The skill may read source files and use local inspection commands to trace security relevant data flow. It does not modify project files, access environment variables, deploy changes, contact external targets, or perform active exploitation.