nsauditor-ai-agent-skill 0.2.20 → 0.2.21

package/CHANGELOG.md

@@ -4,6 +4,12 @@ Release notes for **`nsauditor-ai-agent-skill`** — installable knowledge packa
 
 ---
 
+## 0.2.21 (2026-07-03) — Paired content bump for the CE 0.2.23 operator bug-fix cycle + EE 0.31.9 GAP-1 #3
+
+Paired content bump (SKILL.md version line + this changelog). **CE 0.2.23** hardens cloud-scan scope integrity (a cloud auditor runs only on its own sentinel host — a network `--host` never triggers cloud plugins, including via the MCP `probe_service` tool; `--host` is the sole cloud-intent signal, credentials are not), AI-conclusion robustness (payload-scaled timeout + a fail-visible `scan_response_ai.txt` stub), and the AI bail message. **EE 0.31.9** closes the GAP-1 #3 RoC positive-substrate render-site parity (false-positive controls now render their substrate in HTML too) + FP-correct auditor wording. No new MCP tools or schemas; the `probe_service` tool now enforces the cloud-intent contract (a cloud auditor requires its sentinel host; a network plugin is rejected on a sentinel host). **No new framework, no new plugins (still 28), all seven coverage matrices UNCHANGED.** Paired **CE 0.2.23** + **EE 0.31.9**.
+
+---
+
 ## 0.2.20 (2026-07-01) — Paired content bump for EE 0.31.8 (GAP-1 positive-substrate polish: framework-aware PCI Req 10.5.1 caveat restoration + SOC 2 GRC-push hygiene)
 
 Paired content bump for the EE 0.31.8 cycle (SKILL.md version line + this changelog). EE 0.31.8 restores per-framework precision to the 0.31.7 positive-substrate caveat — the RDS audit-log-retention substrate finding now carries a per-framework caveat override so the PCI report restores the `Req 10.5.1` citation while SOC 2 keeps the neutral base (no PCI leak) — and hardens the mechanism (framework-neutral category so no `pci` machine substring rides the SOC 2 renderJSON → GRC push, empty-string-safe caveat selection, genuine renderJSON GRC-channel tests). **No new framework, no new plugins (still 28), all seven coverage matrices UNCHANGED.** Paired **EE 0.31.8** + CE 0.2.22.

package/SKILL.md

@@ -16,7 +16,7 @@ description: >
 
 # NSAuditor AI — Agent Skill
 
-> **Version:** 0.2.20 (post-EE-0.31.8 — the **GAP-1 positive-substrate polish: framework-aware PCI Req 10.5.1 caveat restoration + SOC 2 GRC-push hygiene** cycle: **no new framework, no new plugins (still 28), all seven coverage matrices UNCHANGED.** A precision-and-hygiene follow-up to 0.31.7's opt-in positive-substrate evidence: the RDS audit-log-retention substrate finding routes to both PCI DSS 10.5.1 and (via inheritance) SOC 2 CC7.2, and now carries a per-framework caveat override so the PCI report restores the `Req 10.5.1` citation while SOC 2 keeps the neutral base — no PCI citation leaks onto a neighbor's RoC; also hardens the mechanism (framework-neutral category so no `pci` machine substring rides the SOC 2 renderJSON → GRC push, empty-string-safe caveat selection, genuine renderJSON GRC-channel tests). The seven frameworks: SOC 2 · HIPAA · NIST CSF 2.0 · PCI DSS v4.0.1 · ISO/IEC 27001:2022 · CIS Controls v8 · **GDPR Article 32** — the last is **Art. 32 infrastructure substrate ONLY, NOT GDPR compliance** [Art. 32 is the only one of GDPR's 99 articles an infrastructure scanner can evidence; the rest is operator-side, out of scope by design], carrying four-factor proportionality + personal-data-scope attestation + the Art. 83(4) LOWER fine tier [€10M/2%, not the €20M/4% headline tier]. Paired CE 0.2.22 + EE 0.31.8) · **Source:** [github.com/nsasoft/nsauditor-ai](https://github.com/nsasoft/nsauditor-ai) · **npm:** `nsauditor-ai` · **License:** MIT (CE)
+> **Version:** 0.2.21 (post-EE-0.31.9 — the **CE 0.2.23 operator bug-fix cycle + EE 0.31.9 GAP-1 #3**: **no new framework, no new plugins (still 28), all seven coverage matrices UNCHANGED.** CE 0.2.23 hardens cloud-scan scope integrity — a cloud auditor runs only on its own sentinel host, so a network `--host` never triggers cloud plugins (including via the MCP `probe_service` tool; `--host` is the sole cloud-intent signal, credentials are not) — plus AI-conclusion robustness (payload-scaled timeout + a fail-visible `scan_response_ai.txt` stub) and the AI bail message; EE 0.31.9 closes the GAP-1 #3 RoC positive-substrate render-site parity (false-positive controls render their substrate in HTML too) + FP-correct auditor wording. The seven frameworks: SOC 2 · HIPAA · NIST CSF 2.0 · PCI DSS v4.0.1 · ISO/IEC 27001:2022 · CIS Controls v8 · **GDPR Article 32** — the last is **Art. 32 infrastructure substrate ONLY, NOT GDPR compliance** [Art. 32 is the only one of GDPR's 99 articles an infrastructure scanner can evidence; the rest is operator-side, out of scope by design], carrying four-factor proportionality + personal-data-scope attestation + the Art. 83(4) LOWER fine tier [€10M/2%, not the €20M/4% headline tier]. Paired CE 0.2.23 + EE 0.31.9) · **Source:** [github.com/nsasoft/nsauditor-ai](https://github.com/nsasoft/nsauditor-ai) · **npm:** `nsauditor-ai` · **License:** MIT (CE)
 
 NSAuditor AI is a modular, AI-assisted network security audit platform with 27+ scanner
 plugins, CVE matching, MITRE ATT&CK mapping, and Zero Data Exfiltration by design. This

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "nsauditor-ai-agent-skill",
-  "version": "0.2.20",
+  "version": "0.2.21",
   "description": "AI Agent Skill for NSAuditor AI — gives any AI coding agent built-in knowledge of NSAuditor's MCP tools, schemas, plugins, and security audit workflows.",
   "keywords": [
     "nsauditor",