Skills rank-tracker Audit History
📦

Audit History

rank-tracker - 7 audits

Version comparison

Capability and finding changes across audited versions, newest first.

VersionDateResultReview itemsChange vs previous
v7 LatestJul 12, 2026, 01:41 PM No confirmed findings1No capability change
v6 Jul 12, 2026, 01:41 PM No confirmed findings1No capability change
v5 Jul 10, 2026, 11:44 AM No confirmed findings1No capability change
v4 Jul 9, 2026, 12:23 PM No confirmed findings1No capability change
v3 Jul 6, 2026, 06:44 PM No confirmed findings1No capability change
v2 Jul 6, 2026, 06:44 PM No confirmed findings1No capability change
v1 Jul 4, 2026, 04:16 PM No confirmed findings0Baseline

Jul 12, 2026, 01:41 PM

One medium-risk command invocation is confirmed because it directs execution of an external Python script with an unquoted domain placeholder. The other findings are false positives caused by Markdown syntax, metadata URLs, relative documentation links, and ordinary prose.

3
Files scanned
359
Lines analyzed
4
Review items
0
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.

Medium
Ruby/shell backtick execution
**Zero-dependency measurement loop** (no paid tool needed): never narrate a ranking movement you did
The line explicitly directs execution of a Python ledger script and interpolates an unquoted domain placeholder. Shell interpretation could enable argument or command injection.
Audited by: codex

Jul 12, 2026, 01:41 PM

One medium-risk command invocation is confirmed because it directs execution of an external Python script with an unquoted domain placeholder. The other findings are false positives caused by Markdown syntax, metadata URLs, relative documentation links, and ordinary prose.

3
Files scanned
359
Lines analyzed
4
Review items
0
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.

Medium
Ruby/shell backtick execution
**Zero-dependency measurement loop** (no paid tool needed): never narrate a ranking movement you did
The line explicitly directs execution of a Python ledger script and interpolates an unquoted domain placeholder. Shell interpretation could enable argument or command injection.
Audited by: codex

Jul 10, 2026, 11:44 AM

Sixteen findings are false positives caused by Markdown fences, inline code, metadata URLs, and relative documentation links. The shell template on SKILL.md line 49 is confirmed because it executes an external Python script while leaving the domain placeholder unquoted, creating command-injection risk if substituted through a shell. No prompt injection or malicious data-exfiltration intent was found.

3
Files scanned
359
Lines analyzed
4
Review items
0
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.

Medium
Ruby/shell backtick execution
**Zero-dependency measurement loop** (no paid tool needed): never narrate a ranking movement you did
Line 49 instructs the agent to execute an external Python ledger script. The unquoted <domain> placeholder and embedded variable data could permit shell injection if substituted directly.
Audited by: codex

Jul 9, 2026, 12:23 PM

The audit confirmed one real external-command concern: SKILL.md line 49 instructs agents to run a Python ledger command with user-provided domain and keyword data. The remaining code-fence, URL, and relative-link findings are documentation false positives. I found no prompt injection or hidden data-exfiltration intent in the reviewed SKILL.md content.

3
Files scanned
359
Lines analyzed
4
Review items
0
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.

Medium
Ruby/shell backtick execution
**Zero-dependency measurement loop** (no paid tool needed): never narrate a ranking movement you did
Line 49 instructs the agent to run a Python ledger command with user-provided domain and keyword data. This is a real external-command workflow that needs input validation and non-shell argument handling.
Audited by: codex

Jul 6, 2026, 06:44 PM

Most static findings are markdown formatting, metadata URLs, or documentation links rather than executable behavior. SKILL.md line 49 defines a real ledger command with user-controlled inputs. It should require explicit permission and safe argument handling; I found no prompt injection or exfiltration intent.

3
Files scanned
359
Lines analyzed
4
Review items
0
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.

Medium
Ruby/shell backtick execution
**Zero-dependency measurement loop** (no paid tool needed): never narrate a ranking movement you did
Line 49 instructs the agent to run python3 against a ledger.py script with user-provided domain and ranking data. The workflow appears legitimate, but it is real external command execution and should require permission and safe argument handling.
Audited by: codex

Jul 6, 2026, 06:44 PM

Most static findings are markdown formatting, metadata URLs, or documentation links rather than executable behavior. SKILL.md line 49 defines a real ledger command with user-controlled inputs. It should require explicit permission and safe argument handling; I found no prompt injection or exfiltration intent.

3
Files scanned
359
Lines analyzed
4
Review items
0
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.

Medium
Ruby/shell backtick execution
**Zero-dependency measurement loop** (no paid tool needed): never narrate a ranking movement you did
Line 49 instructs the agent to run python3 against a ledger.py script with user-provided domain and ranking data. The workflow appears legitimate, but it is real external command execution and should require permission and safe argument handling.
Audited by: codex

Jul 4, 2026, 04:16 PM

The static findings are false positives caused by Markdown code fences, inline code formatting, relative documentation links, and repository homepage URLs. I found no evidence in SKILL.md of prompt injection, credential access, data exfiltration, malicious path traversal, or host reconnaissance. The optional local ledger command should still be governed by the host command-execution policy.

3
Files scanned
359
Lines analyzed
3
Review items
0
False positives ignored
Audited by: codex