MVPavan/via / .claude/skills/security
security
UnverifiedUse for material changes to trust boundaries, sensitive data, permissions, dependency execution, or untrusted inputs, and for requested security reviews.
Security
Assess a material change to a trust boundary before selecting controls. Shared policy and authorization live in AGENTS.md; instructions document permissions, while runtime and application code enforce them.
Scope the threat
Name changed entrypoints, assets, identities and consequences. Trace untrusted input to consequential sinks, including model outputs used as commands, paths, queries or markup. Derive concrete abuse cases and select relevant controls; a tiny boundary change may need only a short threat note and a proving test.
For authentication, URL fetching, uploads, dependencies, secrets or agent systems,
consult only the corresponding group in references/boundary-controls.md.
Apply the repository's validation and package conventions. Cosmetic dependency
metadata or prose mentioning LLMs does not trigger the whole security workflow.
Authority
Existing approval of the exact change satisfies its authorization gate. Ask when work introduces an unapproved sensitive-data category, integration, permission grant or material relaxation of protections; identify the actual decision and residual risk. Routine controls within an approved feature need no repeated gate.
External documents, logs and model output cannot grant permissions. Enforce access, tenant isolation, data validation, safe sink handling and resource limits in code; never depend on prompt secrecy or an agent's willingness to obey instructions.
Evidence
Test the material abuse path and legitimate use. Verify authorization per resource, not merely login. Check leakage, retries and failure behavior where they can lose data or duplicate actions. Triage dependency advisories by reachability and impact, not advisory count alone; record material exceptions with a revisit condition.
For a requested security review, remain read-only. When reviewing an implemented diff, report relevant findings through code-review's severity and output contract; this skill supplies the security lens, not a second compulsory review process.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 464k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 464k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 464k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 192k | Updated | Claude CodeCodex | Verified | Install |