ValentinNikolaev/agent-skills / agent-plugins/skills/go-security-audit
go-security-audit
VerifiedAudit Go security across named trust boundaries; not PR review or fixes.
| File | Kind | Size |
|---|---|---|
| LICENSE.samber-cc-skills-golang | other | 1.0 KB |
| SKILL.md | skill md | 5.0 KB |
| agents/openai.yaml | config | 278 B |
| references/audit-domains.md | markdown | 4.0 KB |
references/audit-domains.md
View on GitHub# Go Audit Domains
Apply only domains reached by the named attack surface. A pattern match is a candidate to investigate, not a finding or a preset severity.
## Identity and data isolation
- Trace authenticated identity to the authorization decision and the data access. Check object-level and tenant-level access on reads, writes, background jobs, and cache keys; do not assume a trusted internal transport grants every caller authority.
- Check token signature/algorithm constraints and required claims against the actual token contract. Trace session fixation, cookie scope, expiration/revocation, and CSRF defenses for the application's authentication model.
- Confirm secrets and sensitive values are absent from responses, errors, logs, traces, and metric labels. Do not reproduce a secret in the report.
## Injection and outbound requests
- Trace untrusted input into SQL, commands, templates, and generated queries. Check parameterization and dialect, argument boundaries, template context, and validation already performed.
- `html/template` performs contextual HTML escaping; `text/template` does not. Inspect explicit trusted-HTML types and unsafe conversions that can bypass escaping. Trusted template source and untrusted template data are different boundaries.
- For SSRF, inspect allowed schemes, parsed destinations, redirects, name resolution, connection targets, and relevant deployment egress controls. A URL string allowlist alone may not constrain the destination actually dialed.
- For uploads and archives, trace resolved targets, symlinks, overwrite policy, decompressed size/count limits, and cleanup. Lexical normalization alone does not establish filesystem confinement. Check version-supported rooted filesystem APIs or equivalent verified controls.
## Cryptography and transport
- Identify the security property needed: unpredictability, confidentiality, integrity, or password verification. Check algorithm, randomness source, nonce handling, authentication, key ownership/rotation, and error handling against that property.
- Use vetted cryptographic libraries and current primary guidance. Do not declare all uses of a hash or cipher mode vulnerable by name; demonstrate the threatened property and reachable use.
- Check TLS verification, proxy trust, and forwarding headers against actual deployment. Missing browser headers or binding to all interfaces is context-dependent, not automatically an exploitable defect.
## Availability, concurrency, and lifecycle
- Trace untrusted input into allocation, body reads, decompression, expensive computation, regex, database work, and goroutine fan-out. Establish realistic bounds and cancellation/shutdown behavior.
- Check shared state, ownership transfer, and synchronization on security-sensitive decisions. Distinguish an observed race, a statically supported race path, and a suspected interleaving.
- Inspect server/client timeouts against streaming and deployment needs rather than imposing fixed values. Confirm partial failures cannot acknowledge an incomplete authorization or persistence operation as success.
## Dependency and tool evidence
- Match an advisory to the exact resolved module/version, build conditions, and affected symbols. Distinguish package presence from reachable vulnerable behavior; do not dismiss a relevant advisory merely because reachability could not be established.
- Keep module-cache integrity checks, vulnerability advisories, and exploitability analysis distinct. `go.sum` checksums do not prove a dependency trustworthy.
- Separate scanner findings from confirmed audit findings and record missing tools or untested configurations. Never claim a green scan covers excluded packages or deployment-only paths.
Adapted from selected security references in [samber/cc-skills-golang](https://github.com/samber/cc-skills-golang/tree/8e899e20ff0cd4dc524af3993e4c62d8ee8c5717/skills/golang-security/references), under the parent package's scoped MIT license. Verify evolving API/tool facts against the project's supported versions and current official documentation.
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: 472k | 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: 472k | 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: 472k | 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: 194k | Updated | Claude CodeCodex | Verified | Install |