sciwrite-process
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions and specific bash snippets for the agent to execute Python scripts (
numeric_consistency.py,keyword_consistency.py,prose_audit.py) intended to audit LaTeX manuscripts for errors and inconsistencies. - [DYNAMIC_EXECUTION]: The skill employs dynamic shell path resolution within its instructions to locate internal scripts. Specifically, it uses shell expansion and directory traversal (
cd -P ~/.claude/skills/sciwrite-prose/../.. && pwd) to calculate the root directory of the toolset at runtime, which is then used to execute Python scripts from computed paths. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process user-provided content in the form of LaTeX files (
paper.tex). - Ingestion points: The manuscript file
paper.texis read and analyzed by multiple audit scripts referenced inSKILL.mdandreferences/final-checklist.md. - Boundary markers: The skill instructions do not specify any boundary markers or delimiters to differentiate between manuscript content and instructions, nor do they include warnings to ignore embedded commands within the analyzed files.
- Capability inventory: The skill allows the agent to execute shell commands and run Python scripts to perform file analysis.
- Sanitization: There is no evidence of sanitization or content filtering for the data ingested from
paper.texbefore it is passed to the execution environment of the audit scripts.
Audit Metadata