apply-grant

Pass

Audited by Gen Agent Trust Hub on Aug 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local bash scripts (export-session.sh) and shell commands to manage session logs and gather git history.
  • [DATA_EXFILTRATION]: Sends telemetry events to a remote Convex endpoint derived from local configuration. This behavior is subject to a user-consent prompt during the initial run.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted data from local project files to generate application drafts. * Ingestion points: Project configuration files (README.md, package.json, Anchor.toml, Cargo.toml) and git history (git log). * Boundary markers: Absent; no delimiters are used to separate ingested data from agent instructions. * Capability inventory: Includes shell command execution and file system read/write access. * Sanitization: Absent; ingested file content is processed directly without filtering or escaping.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 1, 2026, 11:04 AM
Security Audit — agent-trust-hub — apply-grant