Audit History
conversation-memory - 5 audits
Version comparison
Capability and finding changes across audited versions, newest first.
| Version | Date | Result | Review items | Change vs previous |
|---|---|---|---|---|
| v5 Latest | Jul 23, 2026, 09:27 PM | 1 confirmed | 0 | No capability change |
| v4 | Jul 8, 2026, 08:55 AM | No confirmed findings | 0 | No capability change |
| v3 | Jul 6, 2026, 11:24 PM | No confirmed findings | 0 | External commands |
| v2 | Jun 30, 2026, 03:19 PM | No confirmed findings | 0 | No capability change |
| v1 | Feb 24, 2026, 05:38 PM | No confirmed findings | 0 | Baseline |
Jul 23, 2026, 09:27 PM
All 14 static findings are false positives caused by JavaScript template literals, Markdown backticks, and ordinary exception handling. However, examples directly interpolate untrusted messages and recalled memories into prompts, creating a persistent prompt-injection risk.
Confirmed security concerns (1)
Risk Factors
Jul 8, 2026, 08:55 AM
All static findings were false positives caused by JavaScript template literals, markdown code fences, and inline code in SKILL.md. I found no evidence of shell execution, system reconnaissance, prompt-injection text, or malicious intent in the reviewed file.
Risk Factors
Jul 6, 2026, 11:24 PM
Static analysis reported one external command pattern, but the evidence is inline Markdown naming related skills. I found no executable code, shell instructions, network access, data exfiltration, or prompt injection attempts.
Risk Factors
⚙️ External commands (1)
Jun 30, 2026, 03:19 PM
Static analysis flagged Markdown backticks on SKILL.md line 61 as shell execution and front matter on line 3 as weak cryptography. Review found no executable code, no cryptographic implementation, no network access, and no prompt injection attempt. The skill is a guidance-only Markdown file with low residual risk.
Static false positives ignored (2)
These static matches were dismissed by semantic review or matched schema-only tokens, so they are shown for transparency but do not drive the quality score.
Feb 24, 2026, 05:38 PM
Documentation-only skill with no executable code. Static analyzer flagged Markdown backtick formatting as shell commands (false positive) and misidentified description text as cryptographic patterns (false positive). No actual security risks detected. Safe for publication.