The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides instructions to download and execute an installation script from a remote URL using the curl | sh pattern. This is the official installer for the vendor's CLI tool.
Evidence: Found in docs/access-playbook.md and docs/cli-reference.md using the URL https://get.bounded.sh/install.sh.
[COMMAND_EXECUTION]: The documentation includes a Python one-liner to parse and display information from local JSON session files.
Evidence: Found in docs/key-and-account-safety.md: python3 -c "import json;d=json.load(open('$HOME/.bounded/web-session.json'));print(d['email'],'exp',d['expiresAt'])".
[EXTERNAL_DOWNLOADS]: The skill instructs the user to install or update the CLI tool using a package manager.
Evidence: npm install -D @bounded-sh/cli@latest mentioned in docs/access-playbook.md and docs/cli-reference.md.
[CREDENTIALS_UNSAFE]: The skill accesses sensitive local credential files and environment variables used for authentication with the Bounded platform.
Evidence: References to ~/.bounded/credentials, ~/.bounded/web-session.json, and the BOUNDED_PRIVATE_KEY environment variable appear throughout the documentation (e.g., docs/key-and-account-safety.md).
[INDIRECT_PROMPT_INJECTION]: The skill defines a workflow that processes data from external sources while having the authority to perform administrative actions.
Ingestion points: Data is ingested through commands such as bounded subscribe, bounded data get, and bounded data search (documented in docs/cli-reference.md).
Boundary markers: No explicit delimiters or instruction-ignore warnings are specified for the processed data stream.
Capability inventory: The skill includes powerful deployment and configuration capabilities via bounded deploy, bounded site deploy, and bounded functions deploy.
Sanitization: The skill relies on the platform's server-side policy enforcement for sanitization and validation of operations.

bounded-deploy