The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructions in SKILL.md explicitly command the agent to 'NOT further summarize or alter the content' of the API response and to 'return the formatted markdown' exactly as provided. This creates a high susceptibility to indirect prompt injection if the get-tldr.com API processes a malicious webpage containing instructions designed to hijack the agent's behavior.
[DATA_EXFILTRATION]: The script get_tldr.py performs network operations to https://www.get-tldr.com/api/v1/summarize. While this is the primary purpose of the skill, it transmits a user-provided URL to an external service. No sensitive local data access was detected.
[COMMAND_EXECUTION]: The skill executes a bundled Python script get_tldr.py. The script uses the requests library to make HTTP calls and logs the activity to a local file (~/.config/get-tldr/skill.log). While it performs file writes and network requests, these are confined to the intended functionality of the tool.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: Data enters the agent via the get_tldr.py script which fetches content from https://www.get-tldr.com/api/v1/summarize.
Boundary markers: None identified. The instructions specifically tell the agent to treat the output as authoritative and not to modify it.
Capability inventory: The skill uses requests for network access and open() for appending to a log file in the user's home directory.
Sanitization: There is no evidence of sanitization or filtering of the API response before it is presented to the agent.

get-tldr