The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it processes untrusted data from external platforms.
Ingestion points: manifest.json and SKILL.md workflow steps indicate the ingestion of news sentiment, earnings transcripts, Twitter chatter, and Reddit comments.
Boundary markers: The instructions lack specifications for using delimiters or protective instructions to isolate untrusted external content from the agent's core logic.
Capability inventory: The skill facilitates multiple network requests via a manifest runner to fetch external data.
Sanitization: No evidence of sanitization, filtering, or validation of the ingested social media or news content is provided.
[COMMAND_EXECUTION]: The skill instructions rely on the execution of local Python scripts to process the manifest.
Evidence: SKILL.md identifies ../selat-purchasing/scripts/run_manifest.py and ../selat-purchasing/scripts/spend_report.py as the primary execution and reporting tools.
[DATA_EXFILTRATION]: The skill communicates with several external domains that are not on the standard whitelist.
Domains: alphavantage.mpp.paywithlocus.com, stableenrich.dev, and x402.ottoai.services.
Context: These network operations are intended for retrieving financial data from designated providers. The domain catalog.selat.ai is identified as a vendor-owned resource.

stock-direction-signals