The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from the Clari platform and formats it into Slack notifications. Malicious or malformed data within Clari could potentially influence the content of these notifications.
Ingestion points: Data is retrieved from the Clari API in SKILL.md (Step 3) using client.export_and_download.
Boundary markers: The skill does not use specific delimiters or instructions to prevent the agent from interpreting embedded content in the forecast data.
Capability inventory: The skill uses requests.post for network communication (Slack) and open() for local file writes (data/latest.json).
Sanitization: There is no evidence of HTML escaping or data sanitization before the forecast information is placed into Slack message blocks.
[DYNAMIC_EXECUTION]: The skill provides a bash script that executes a Python code block via a command-line flag.
Evidence: In Step 3, the clari-monitor.sh template uses python3 -c to execute logic for change detection and alerting.
[COMMAND_EXECUTION]: The skill documentation includes instructions for executing shell-based commands for automation and API interaction.
Evidence: Usage of curl for webhook registration and bash for scheduled monitoring tasks.

clari-webhooks-events