The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent or user to download and execute code from a third-party, non-trusted GitHub repository. This is an unverified dependency that executes arbitrary code during setup.
Evidence: Instructions to run 'git clone https://github.com/jamiepine/voicebox.git' followed by 'make setup' in 'references/generative-tools.md'.
[COMMAND_EXECUTION]: The skill relies on local CLI scripts to interact with sensitive advertising platform APIs for reporting and campaign management.
Evidence: Execution patterns such as 'node tools/clis/google-ads.js' and various campaign management commands documented in 'SKILL.md'.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection through its core functionality of processing external performance data.
Ingestion points: Performance data provided via CSV files, text pastes, or API outputs as described in the iteration workflow of 'SKILL.md'.
Boundary markers: Absent. There are no instructions or delimiters defined to prevent the agent from following instructions embedded within the untrusted performance data.
Capability inventory: The agent has access to tools capable of creating and modifying paid campaigns across Google, Meta, LinkedIn, and TikTok platforms.
Sanitization: Absent. No evidence of sanitization or validation of the input data is provided.

ad-creative