📦

Audit History

python-performance-optimization - 8 audits

Version comparison

Capability and finding changes across audited versions, newest first.

VersionDateResultReview itemsChange vs previous
v8 LatestJul 8, 2026, 01:34 PM No confirmed findings0No capability change
v7 Jul 8, 2026, 01:34 PM No confirmed findings0 Filesystem access
v6 Jul 1, 2026, 12:25 AM No confirmed findings3No capability change
v5 Jul 1, 2026, 12:25 AM No confirmed findings3Filesystem access
v4 Jan 17, 2026, 09:26 AM No confirmed findings0No capability change
v3 Jan 17, 2026, 09:26 AM No confirmed findings0 Filesystem access
v2 Jan 4, 2026, 05:01 PM No confirmed findings3No capability change
v1 Jan 4, 2026, 05:01 PM No confirmed findings3Baseline

Jul 1, 2026, 12:25 AM

Static analysis reported many command-execution alerts, but review shows they are Markdown code fences and example profiler commands, not executable skill code. The skill is documentation-only, with low residual risk from copied examples that run local profilers, open local files, or call httpbin.org for I/O benchmarking.

1
Files scanned
870
Lines analyzed
6
Review items
1
False positives ignored
Capability review items (3)

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
Example Commands Profile Local Processes
The py-spy examples inspect a local process by PID and can expose stack information if a user runs them against sensitive production services. This is legitimate performance tooling, and no evidence shows the skill executes these commands automatically.
The line-numbered context shows explicit py-spy commands against a PID. The risk depends on a user manually running the documented command, so severity remains low.
Low
Example HTTP Benchmark Uses External Service
The async I/O example imports requests and aiohttp and calls https://httpbin.org/delay/1. This is a benign timing demonstration, but copied code would make outbound network requests.
The URLs and HTTP client calls are visible in the referenced lines. They target a public test endpoint and do not include credentials or exfiltration logic.
Low
Examples Open Local Files and Databases
The SQLite and iterator examples open example.db or a caller-provided filename. This is normal optimization guidance, but users should test with non-sensitive data.
The examples clearly demonstrate local file and database access. They are not hidden behavior because SKILL.md is the only file and contains documentation examples only.
Static false positives ignored (1)

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
Static Command Alerts Are Markdown False Positives
The many Ruby or shell backtick findings correspond to Markdown fenced code blocks and sample snippets. No standalone script, installer, or automatic command runner was found in the skill package.
Line-numbered review shows fenced code examples rather than executable Ruby backticks. The package contains only SKILL.md, so there is no execution surface outside user-copied examples.
Audited by: codex

Jul 1, 2026, 12:25 AM

Static analysis reported many command-execution alerts, but review shows they are Markdown code fences and example profiler commands, not executable skill code. The skill is documentation-only, with low residual risk from copied examples that run local profilers, open local files, or call httpbin.org for I/O benchmarking.

1
Files scanned
870
Lines analyzed
6
Review items
1
False positives ignored
Capability review items (3)

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
Example Commands Profile Local Processes
The py-spy examples inspect a local process by PID and can expose stack information if a user runs them against sensitive production services. This is legitimate performance tooling, and no evidence shows the skill executes these commands automatically.
The line-numbered context shows explicit py-spy commands against a PID. The risk depends on a user manually running the documented command, so severity remains low.
Low
Example HTTP Benchmark Uses External Service
The async I/O example imports requests and aiohttp and calls https://httpbin.org/delay/1. This is a benign timing demonstration, but copied code would make outbound network requests.
The URLs and HTTP client calls are visible in the referenced lines. They target a public test endpoint and do not include credentials or exfiltration logic.
Low
Examples Open Local Files and Databases
The SQLite and iterator examples open example.db or a caller-provided filename. This is normal optimization guidance, but users should test with non-sensitive data.
The examples clearly demonstrate local file and database access. They are not hidden behavior because SKILL.md is the only file and contains documentation examples only.
Static false positives ignored (1)

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
Static Command Alerts Are Markdown False Positives
The many Ruby or shell backtick findings correspond to Markdown fenced code blocks and sample snippets. No standalone script, installer, or automatic command runner was found in the skill package.
Line-numbered review shows fenced code examples rather than executable Ruby backticks. The package contains only SKILL.md, so there is no execution surface outside user-copied examples.
Audited by: codex

Jan 4, 2026, 05:01 PM

Documentation-only skill containing Python optimization examples. No executable scripts, network calls, or file access capabilities. Users receive guidance only; no code is executed on their system.

4
Files scanned
1,084
Lines analyzed
6
Review items
0
False positives ignored
Capability review items (3)

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
Network request examples in docs
The guide includes sample HTTP calls such as `requests.get(url)` and endpoints like `https://httpbin.org/delay/1`. These snippets are documentation only; they are not executed by the skill and pose no risk to users.
Low
Shell command execution examples
The documentation shows commands like `python -m cProfile -o output.prof script.py`. These are instructional examples for users to run manually, not automatic executions by the skill.
Low
Local file access examples
Examples demonstrate `with open(filename) as f:` for processing files. Users would run these patterns manually if needed; the skill itself does not access any files.

Risk Factors

🌐 Network access (1)
⚙️ External commands (1)
📁 Filesystem access (1)
Audited by: claude

Jan 4, 2026, 05:01 PM

Documentation-only skill containing Python optimization examples. No executable scripts, network calls, or file access capabilities. Users receive guidance only; no code is executed on their system.

4
Files scanned
1,084
Lines analyzed
6
Review items
0
False positives ignored
Capability review items (3)

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
Network request examples in docs
The guide includes sample HTTP calls such as `requests.get(url)` and endpoints like `https://httpbin.org/delay/1`. These snippets are documentation only; they are not executed by the skill and pose no risk to users.
Low
Shell command execution examples
The documentation shows commands like `python -m cProfile -o output.prof script.py`. These are instructional examples for users to run manually, not automatic executions by the skill.
Low
Local file access examples
Examples demonstrate `with open(filename) as f:` for processing files. Users would run these patterns manually if needed; the skill itself does not access any files.

Risk Factors

🌐 Network access (1)
⚙️ External commands (1)
📁 Filesystem access (1)
Audited by: claude