The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions in the README recommend installing the autark or autark-cli package globally via the NPM registry.
[COMMAND_EXECUTION]: The skill primarily functions by instructing the agent to execute shell commands using the autark CLI to manage data models, record outreach, and start/finish work runs.
[CREDENTIALS_UNSAFE]: The CLI tool manages and accesses authentication credentials locally at ~/.autark/credentials.json, which the skill relies on for platform interactions.
[DATA_EXFILTRATION]: The skill sends collected market research data, including prospect names, emails, and bios, to the platform's API at autark-api.kushalsokke.workers.dev to populate the market-discovery loop.
[REMOTE_CODE_EXECUTION]: The documentation includes an autark update command designed to download and execute updates for the CLI tool from the vendor's remote infrastructure.
[PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes data from untrusted external sources like web pages and repositories.
Ingestion points: Sourcing loop (crawling sites, reading repositories, polling email threads) described in SKILL.md.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands in external data are present.
Capability inventory: Subprocess execution of the autark CLI, local file writing, and network operations for email outreach.
Sanitization: No sanitization or validation of the ingested external content is mentioned before it is processed or recorded.

autark