Skills
Navigate
skills/ohad6k/emulo/agent-profile-from-sessions/Gen Agent Trust Hub

agent-profile-from-sessions

Pass

Audited by Gen Agent Trust Hub on Oct 2, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill uses uvx to download and execute the emulo package (version 0.6.7) and npx to install an additional skill (ohad6k/emulo@emulo). Both represent tools provided by the skill author.
  • [COMMAND_EXECUTION]: The skill executes the emulo command-line tool with the --coach flag and specific source arguments (e.g., --source claude) to process local files.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-typed messages from past sessions which may contain embedded instructions or adversarial content.
  • Ingestion points: Reads local session logs from Claude Code, Codex, Copilot CLI, OpenCode, and Google Antigravity (SKILL.md).
  • Boundary markers: None explicitly defined for the analysis phase; however, it specifies mining only user messages.
  • Capability inventory: Executes shell commands via uvx and performs file writes for profile installation (SKILL.md).
  • Sanitization: Instructions require the user to manually review every quote in the generated profile and remove unwanted content before installation.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 2, 2026, 12:13 PM