dailybot

Pass

Audited by Gen Agent Trust Hub on May 29, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads its official CLI installer script and associated verification assets (SHA-256 checksums and OIDC certificates) from the vendor's official domain (cli.dailybot.com) and GitHub organization.
  • [REMOTE_CODE_EXECUTION]: Installation involves piping remote scripts to the system shell (bash or iex). However, this behavior is mitigated by documented defense-in-depth integrity checks, including mandatory SHA-256 verification and cross-origin comparison against the public GitHub source before execution.
  • [COMMAND_EXECUTION]: The skill makes extensive use of local shell commands to drive the dailybot CLI tool and uses git to perform environment detection for repository and branch metadata.
  • [DATA_EXFILTRATION]: The dailybot-email capability provides an outbound data path. This risk is managed through a mandatory recipient approval cache (~/.dailybot/email-approvals.json) and an automated regex-based scanner that identifies and blocks the transmission of sensitive data such as AWS keys, Stripe tokens, GitHub PATs, and private keys.
  • [PROMPT_INJECTION]: The skill incorporates a robust trust model for untrusted input, such as team messages and form responses. It explicitly instructs the agent to treat these as advisory and requires explicit developer confirmation before performing any side-effecting actions derived from message content.
Audit Metadata
Risk Level
SAFE
Analyzed
May 29, 2026, 08:26 PM
Security Audit — agent-trust-hub — dailybot