The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill retrieves data from external datasets and incorporates it into reports. This creates a potential surface for indirect prompt injection if an attacker were to place malicious instructions within the source data (e.g., within permit application fields in the public NYC database).
Ingestion points: JSON data retrieved from data.cityofnewyork.us via WebFetch in SKILL.md.
Boundary markers: Absent; data is directly interpolated into Markdown tables without explicit delimiters or warnings to ignore embedded instructions.
Capability inventory: Access to Write (file system), Bash (shell execution), and WebFetch (network requests) as defined in the allowed-tools section of SKILL.md.
Sanitization: Absent; the instructions do not specify any validation, filtering, or escaping of the external data before it is presented or saved.
[EXTERNAL_DOWNLOADS]: The skill fetches property metadata and permit records from the official City of New York Open Data portal.
Source: Fetches records from data.cityofnewyork.us/resource/ using WebFetch as described in the Step 3 section of SKILL.md.

nyc-dob-permits