📦

Audit History

social-media-analyzer - 6 audits

Version comparison

Capability and finding changes across audited versions, newest first.

VersionDateResultReview itemsChange vs previous
v6 LatestJul 6, 2026, 03:09 AM 1 confirmed0No capability change
v5 Jul 6, 2026, 03:09 AM 1 confirmed0External commands Contains scripts
v4 Jun 28, 2026, 09:02 AM No confirmed findings1Contains scripts Filesystem access
v3 Jan 16, 2026, 02:56 PM No confirmed findings0No capability change
v2 Jan 16, 2026, 02:56 PM No confirmed findings0Filesystem access External commands
v1 Jan 15, 2026, 11:55 AM No confirmed findings0Baseline

Jul 6, 2026, 03:09 AM

The markdown backtick and system reconnaissance static findings are false positives from documentation and generated text, not executable behavior. A separate high-severity semantic issue remains because eval_result.json embeds self-declared safe-to-publish audit claims that could bias automated review.

7
Files scanned
740
Lines analyzed
2
Review items
0
False positives ignored

Confirmed security concerns (1)

High
Prompt Injection Attempt Detected
eval_result.json declares "risk_level" as "safe", sets "safe_to_publish" to true, and states "All 8 static findings are FALSE POSITIVES." This embedded audit verdict could bias automated review instead of providing neutral skill instructions.
Lines 15-18 contain self-declared publication safety and a broad false-positive claim. It is not a direct override command, so confidence is high but not maximal.
Audited by: codex

Jul 6, 2026, 03:09 AM

The markdown backtick and system reconnaissance static findings are false positives from documentation and generated text, not executable behavior. A separate high-severity semantic issue remains because eval_result.json embeds self-declared safe-to-publish audit claims that could bias automated review.

7
Files scanned
740
Lines analyzed
2
Review items
0
False positives ignored

Confirmed security concerns (1)

High
Prompt Injection Attempt Detected
eval_result.json declares "risk_level" as "safe", sets "safe_to_publish" to true, and states "All 8 static findings are FALSE POSITIVES." This embedded audit verdict could bias automated review instead of providing neutral skill instructions.
Lines 15-18 contain self-declared publication safety and a broad false-positive claim. It is not a direct override command, so confidence is high but not maximal.
Audited by: codex

Jun 28, 2026, 09:02 AM

AI review dismissed the static weak-crypto, shell-backtick, and system-reconnaissance findings as false positives in documentation, JSON text, and Python docstrings. The skill contains local Python analytics scripts, but reviewed logic is limited to metric calculations, benchmark comparison, and recommendations with no network, filesystem write, subprocess, environment, or credential access.

7
Files scanned
740
Lines analyzed
2
Review items
3
False positives ignored
Capability review items (1)

These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.

Low
Local Python Analytics Scripts
The skill includes Python scripts that calculate metrics and recommendations from provided campaign data. Review found local arithmetic and benchmark logic only, with no file writes, network calls, subprocess execution, environment access, or credential handling.
The executable files are present and review confirms their purpose. Confidence is high because the relevant modules were inspected, but it remains a community skill with executable code.
Static false positives ignored (3)

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.

Low
False Positive: Weak Cryptographic Algorithm Matches
Static analysis flagged weak cryptographic algorithm patterns at these locations. Review found docstrings, prose, JSON report text, and field values; no cryptographic API, hash function, or security-sensitive algorithm is present.
The cited Python lines are module docstrings and the cited Markdown and JSON lines are descriptive text. I found no evidence of cryptographic code or weak algorithms at these locations.
Low
False Positive: Shell Backtick Execution Matches
Static analysis flagged Ruby or shell backtick execution at these locations. The reviewed lines use backticks as Markdown inline-code formatting or as JSON text copied from a prior report, not executable shell syntax.
The flagged SKILL.md lines document script filenames in Markdown. The eval_result.json lines are inert JSON strings and are not executed by the skill.
Low
False Positive: System Reconnaissance Matches
Static analysis flagged system reconnaissance language. The cited text discusses marketing platform benchmarks and campaign comparison, with no commands or logic that enumerate the host system, users, processes, or network.
The phrase is business analytics guidance, not operating system reconnaissance. No evidence of host inspection or data collection appears in the reviewed context.

Risk Factors

Audited by: codex

Jan 16, 2026, 02:56 PM

All 26 static findings are false positives. Scanner misidentified docstrings as crypto code, markdown backticks as shell execution, and marketing terminology as system reconnaissance. Code contains only benign Python arithmetic for engagement and ROI calculations.

8
Files scanned
977
Lines analyzed
1
Review items
0
False positives ignored

Risk Factors

Audited by: claude

Jan 16, 2026, 02:56 PM

All 26 static findings are false positives. Scanner misidentified docstrings as crypto code, markdown backticks as shell execution, and marketing terminology as system reconnaissance. Code contains only benign Python arithmetic for engagement and ROI calculations.

8
Files scanned
977
Lines analyzed
1
Review items
0
False positives ignored

Risk Factors

Audited by: claude

Jan 15, 2026, 11:55 AM

All 8 static findings are false positives. The scanner misidentified markdown code formatting as shell execution and benign documentation text containing 'ROI' as cryptographic patterns. The codebase contains only mathematical calculations for engagement metrics and no dangerous patterns.

6
Files scanned
545
Lines analyzed
1
Review items
0
False positives ignored

Risk Factors

⚙️ External commands (2)
Audited by: claude