The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from external Linear tickets (titles, descriptions, and comments). This content is considered untrusted as it originates from outside the skill's instructions. An attacker could place malicious instructions inside a ticket to try and influence the agent's behavior during the estimation process.
Ingestion points: Data fetched via the lc issue list command in Step 1.
Boundary markers: The skill does not provide specific instructions to the agent on how to delimit or ignore potentially malicious instructions embedded within the ticket data.
Capability inventory: The agent has the ability to execute CLI commands (lc) and write files to the local file system (~/work/estimates/index.html).
Sanitization: There are no instructions to sanitize or escape the ticket content before it is rendered into the HTML report, which could lead to minor cross-site scripting (XSS) issues in the generated report if viewed in a browser.
[COMMAND_EXECUTION]: The skill uses the lc (Linear CLI) tool to fetch ticket data and mkdir to create directories. These are standard operations for the skill's functionality and do not exhibit suspicious patterns.

estimate-all