resume-claude-here
Pass
Audited by Gen Agent Trust Hub on Apr 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Python helper script (
scripts/claude_session_tool.py) and the officialclaudeCLI tool to manage session history. These commands are used to discover, rank, and export local JSONL session logs. - [PROMPT_INJECTION]: The skill facilitates the ingestion of data from past conversation transcripts, creating a surface for indirect prompt injection if those transcripts contain malicious instructions from previous sessions.
- Ingestion points: Reads local session files from the
~/.claude/directory. - Boundary markers: The instructions do not prescribe specific delimiters or instruction-ignore warnings when importing transcript text.
- Capability inventory: Includes subprocess execution of the helper script and the
claudeCLI, as well as local file system writes. - Sanitization: Imported transcript content is brought into the context without specific filtering or sanitization of embedded commands.
Audit Metadata