The Agent Skills Directory

[PRIVILEGE_ESCALATION]: The skill includes functionality to modify sensitive system configuration files and credential stores. Evidence: In scripts/agent.py, the create_deployment_plan function lists target paths such as /etc/app/config.yml and /root/.ssh/config. Modifying files in the root directory or global configuration directories typically requires elevated permissions and could be misused if not properly constrained.
[COMMAND_EXECUTION]: The skill performs direct file system modifications across multiple sensitive paths. Evidence: The deploy_aws_credential_token function creates and writes to /opt/backup/.aws/credentials, and deploy_dns_token_in_config appends configuration entries to files at paths like /etc/app/config.yml. These operations involve writing structured data directly to the file system.
[EXTERNAL_DOWNLOADS]: The skill makes network requests to generate security tokens from an external service. Evidence: The scripts/agent.py file uses the requests library to send POST requests to https://canarytokens.org/generate. This is a well-known service in the security community for generating breach detection tokens, consistent with the skill's stated purpose.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external API responses and local log files, creating a vulnerability surface. Evidence: 1. Ingestion points: The functions create_dns_canarytoken, create_web_bug_token, and create_aws_key_token in scripts/agent.py process JSON responses from the Canarytokens API. The check_token_alerts function reads and parses JSON from local log files. 2. Boundary markers: There are no delimiters or instructions to ignore embedded commands when processing these inputs. 3. Capability inventory: The skill can write files, create directories, and perform network operations. 4. Sanitization: While deploy_database_honeytoken validates table names, there is no validation or sanitization of data retrieved from the external API or logs before it is processed.

implementing-honeytokens-for-breach-detection