The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches run metadata and test artifacts (screenshots, DOM snapshots, logs) from the official API endpoint at api.momentic.ai. These assets are downloaded as ZIP archives and extracted locally for analysis.
[COMMAND_EXECUTION]: The skill utilizes the momentic CLI tool via npx for result classification and automated triage. It also provides shell script templates that use curl, jq, and unzip to manage local artifact inspection. The ai triage command specifically supports a --start flag which allows users to execute local build and start commands as part of the test repair process.
[INDIRECT_PROMPT_INJECTION]: The skill implements a workflow where AI agents analyze data from external applications, such as DOM content, console logs, and network traces. This ingestion of untrusted external content represents a potential surface for indirect prompt injection targeting the test-repair logic.
Ingestion points: Test execution artifacts (DOM, logs, screenshots) processed by momentic_get_step_result and momentic ai triage (SKILL.md).
Boundary markers: The instructions do not specify explicit delimiters for the ingested content, relying on the underlying momentic CLI to manage the agent context.
Capability inventory: The skill has the ability to execute shell commands (via npx and --start), perform network operations (via curl), and modify the local filesystem (via unzip and ai triage edits).
Sanitization: The documentation does not describe explicit sanitization or filtering of the processed web application content.

momentic-maintain