Skills
skills/kaelzhang321/teambition-agent-toolkit/tb-get-my-project-tasks/Gen Agent Trust Hub

tb-get-my-project-tasks

Warn

Audited by Gen Agent Trust Hub on Apr 18, 2026

Risk Level: MEDIUMCREDENTIALS_UNSAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill instructions in SKILL.md explicitly direct the agent to use the Read tool on the .teambition.md configuration file to extract an Operator ID. However, this file also contains the App Secret and App ID (as utilized by the tb-api.mjs script), meaning these sensitive credentials are exposed to the AI agent's prompt context when the file is read.
  • [COMMAND_EXECUTION]: The skill executes a local Node.js script scripts/tb-api.mjs via the Bash tool. The execution path includes the projectId parameter, which is obtained via user interaction through AskUserQuestion.
  • [DATA_EXFILTRATION]: The script scripts/tb-api.mjs performs network requests using fetch to open.teambition.com, a well-known collaboration service. These requests are used to exchange the App Secret for an access token and to retrieve project tasks.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it retrieves and displays task content from the TeamBition API. Ingestion points: Task titles and notes are fetched from open.teambition.com via the get-project-tasks command. Boundary markers: The skill does not employ delimiters or specific instructions to ignore embedded commands in the retrieved data. Capability inventory: The agent has access to Bash (shell execution) and Read (file access). Sanitization: No validation or sanitization is performed on the API-returned content before it is displayed to the user in a table.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 18, 2026, 02:21 PM
Security Audit — agent-trust-hub — tb-get-my-project-tasks