The Agent Skills Directory

[Prompt Injection] (MEDIUM): The skill is vulnerable to Indirect Prompt Injection (Category 8) because it ingests untrusted external data and uses it to influence agent output.
Ingestion points: scripts/scrape_my_posts.py retrieves post content from the Apify API.
Boundary markers: Absent. The format_posts_for_analysis function uses weak --- POST N --- delimiters which are easily bypassed by embedded instructions.
Capability inventory: The agent uses ingested data to write new LinkedIn posts, which could be manipulated to include malicious content or exfiltration links.
Sanitization: Absent. No sanitization or escaping of the scraped content is performed before it is provided to the LLM.
[External Downloads] (LOW): The script scripts/scrape_my_posts.py makes network requests to api.apify.com using urllib.request. While required for functionality, this domain is not on the Trusted External Sources list.
[Credentials Unsafe] (LOW): The skill requires an APIFY_TOKEN to be stored in the user's shell environment. While this is standard practice for CLI tools, it increases the risk of credential exposure if the environment is compromised.

writing-linkedin-posts