Perform a dedicated nanace security review for authentication, authorization/RBAC, RLS and tenant isolation, Qdrant filtering, secrets, MCP exposure, Google sync, or other security boundaries. Use only when the user requests a security review or an adversarial review explicitly escalates to the full threat map; do not auto-load solely because code touches a sensitive path.
Installs into .claude/skills of the current project.
Sponsors
SSkills DirectorySkills Directory
Is that skill safe to install?
SSkills DirectorySkills Directory
Know which skills are safe — weekly.
Are you the author of Security Advisory Lessons?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/felichita-security-advisory-lessons)
---
name: security-advisory-lessons
description: Perform a dedicated nanace security review for authentication, authorization/RBAC, RLS and tenant isolation, Qdrant filtering, secrets, MCP exposure, Google sync, or other security boundaries. Use only when the user requests a security review or an adversarial review explicitly escalates to the full threat map; do not auto-load solely because code touches a sensitive path.
---
# nanace Security Review
Use this skill as the deep security lens. For a normal adversarial review with a
matched security surface, the concise security reference under
`adversarial-validation` is sufficient.
## Workflow
1. Freeze the exact diff/head and identify the changed trust boundaries.
2. Inspect the headings in [threat-patterns.md](references/threat-patterns.md),
then read the matching sections. Read the full map only for a broad audit.
3. Trace unauthenticated, outsider (other org), wrong-team member,
low-role member (reader), removed member, malformed-input, and
default-config cases. Security decisions must fail closed.
4. Require a focused negative regression test for the bypass form (HTTP-level,
plus a `scoped_tx` test when tenancy is involved), not only the intended
success path. State residual risk when a test is impractical.
5. Report proven vulnerabilities separately from defense-in-depth hardening.
## Finding Standard
Every candidate names the lower-trust principal, the boundary it crosses, the
affected resource, and the observable result. A best-practice gap or an impact
confined to the caller's own principal is hardening, not a candidate.
Report each candidate under exactly one verdict:
- `confirmed`: a complete `file:line` source trace and a bounded observed
result. Only confirmed findings carry severity; each also states attacker
prerequisites, concrete input/path, impact, smallest safe fix, and a
regression check.
- `needs_validation`: the trace is blocked on one exact unresolved fact, such
as deployment or infrastructure behavior. State that fact; assign no severity
and do not substitute an assumption.
- `rejected`: a disproved candidate, with the caller, invariant, or test that
disproves it.
Do not exaggerate unauthenticated impact when the actual issue requires
authenticated low privilege.