📦
Audit History
fix-sentry-issues - 1 audit
Audit version 1 Latest
Aug 21, 2026, 08:34 AM
Most backtick detections are Markdown code formatting, not Ruby or shell execution. The skill directs agents to run Axiom and Git commands and fetch production-derived URLs, creating command, network, and data-handling risks. No prompt injection or malicious intent was found.
1
Files scanned
73
Lines analyzed
6
Review items
0
False positives ignored
Confirmed security concerns (1)
Medium
Production Data Reuse Without Redaction
The workflow directs agents to reproduce failures and create tests with actual Sentry inputs without requiring secret or personal-data redaction.
The instructions explicitly require actual production inputs in reproduction and tests. Sentry events can contain sensitive values, while no redaction step is specified.
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.
Medium
Ruby/shell backtick execution
2. **Cross-reference Axiom** — Events have `traceId`. `axiom query "['shiori-events'] | where traceI
The line explicitly directs the agent to run the Axiom CLI through a shell command. It interpolates a trace identifier and contacts an external logging service.
Medium
Ruby/shell backtick execution
One branch per issue. `git checkout main && git pull && git checkout -b fix/<descriptive-name>`
The line explicitly instructs execution of chained Git shell commands that change branches and pull from a remote. The generated branch placeholder requires validation.
Medium
Fetch API call
5. **Reproduce** — Use actual failing inputs from Sentry. Call the function with exact data. `fetch(
The workflow explicitly instructs fetching URLs copied from production Sentry events. Unvalidated event-controlled URLs can reach untrusted or private network targets.
Risk Factors
⚙️ External commands (8)
🌐 Network access (1)
Audited by: codex