Skip to content

Melvynx/saveit.now / .agents/skills/ns-security-audit

ns-security-audit

Verified

Audit and remediate NowStack Mobile security flaws. Use for auth/session abuse, Apple or Stripe payment bypass, Convex authorization, cross-user data leaks, unsafe deep links, R2 uploads, client secrets, SSRF, XSS, or pentest follow-up.

FileKindSize
SKILL.mdskill md7.6 KB
agents/openai.yamlconfig368 B
assets/codex-icon.svgother398 B
references/audit-controls.mdmarkdown5.3 KB
references/evidence-and-reporting.mdmarkdown2.8 KB
scripts/security-audit-scan.shexecutablescript2.5 KB

references/evidence-and-reporting.md

View on GitHub
# Evidence And Reporting

## Evidence Levels

Use the strongest level actually achieved:

1. `static`: reachable code path and missing/present guard inspected;
2. `automated`: focused regression test exercises attacker and valid paths;
3. `local runtime`: behavior observed against local web/Simulator and isolated
   data;
4. `staging`: behavior observed on an explicitly authorized staging target;
5. `production`: low-impact behavior observed on an explicitly authorized
   production target or exact distributed binary.

Never promote one level into another. A build passing is not runtime proof, a
Simulator is not a store binary, and staging proof is not production proof.

## Finding Status

- `CONFIRMED`: reachable attacker-controlled path and security impact proven.
- `FIXED`: confirmed root cause repaired and regression evidence passed.
- `CANDIDATE`: suspicious code or configuration requiring more evidence.
- `NOT REPRODUCED`: an authorized test did not reproduce the claimed behavior.
- `NOT VERIFIABLE`: safe proof is unavailable or would cross a boundary.

Do not report the absence of a visible guard as confirmed when a framework,
provider, platform entitlement, proxy, or shared helper may supply the control.

## Severity Calibration

- `Critical`: practical unauthenticated takeover, broad sensitive data access,
  or reliable payment/entitlement bypass with major impact.
- `High`: substantial privilege, paid-feature, stored-XSS, identity, or token
  compromise requiring limited prerequisites.
- `Medium`: meaningful scoped exposure, cost abuse, missing quota, constrained
  SSRF, unsafe link/WebView path, or auth weakness with practical prerequisites.
- `Low`: limited disclosure or defense-in-depth gap.
- `Info`: secure behavior, quality issue, or observation without impact.

Severity combines exploitability, affected users, data sensitivity, financial
impact, persistence, platform protections, and compensating controls.

## Finding Template

```markdown
### [SEVERITY] Short title — STATUS

- Boundary: component and trust boundary, without live secrets or PII
- Impact: what an attacker could achieve
- Evidence: redacted file/line, test, or runtime observation
- Root cause: trusted value or invariant that was missing
- Remediation: shared enforcement point and regression coverage
- Verification: static | automated | local runtime | staging | production
- Residual risk: remaining limitation or NOT VERIFIABLE evidence
```

Also list tested secure boundaries. This prevents future audits from converting
known negative checks into findings and makes partial coverage explicit.

## Redaction Rules

Never include session/recovery tokens, cookies, receipts, API keys, webhook
secrets, full provider IDs, push tokens, personal data, private URLs, complete
database rows, or reusable exploit payloads. Use placeholders and describe the
invariant.
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: 471kUpdated Claude CodeCodexVerifiedInstall
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: 471kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 194kUpdated Claude CodeCodexVerifiedInstall