The Agent Skills Directory

[SAFE]: No security issues were identified. The skill performs local data processing and reporting as described in its documentation.
[DATA_EXFILTRATION]: The skill reads data from local files within a user-provided directory. It does not access sensitive system files or attempt network exfiltration.
[COMMAND_EXECUTION]: The aggregation script uses yaml.safe_load() for secure data ingestion. It contains no instances of eval(), exec(), or subprocess calls.
[REMOTE_CODE_EXECUTION]: No network activity or external downloads were detected in the source code or instructions.
[PROMPT_INJECTION]: The skill handles data from external files, creating a potential surface for indirect prompt injection. This is evaluated as safe due to the absence of dangerous capabilities. 1. Ingestion points: External data enters through load_json_files and load_yaml_files in scripts/aggregate_signals.py via command-line glob patterns. 2. Boundary markers: Absent; the skill does not wrap processed content in delimiters when generating reports. 3. Capability inventory: Capabilities are limited to file-writing in scripts/aggregate_signals.py; no subprocess, exec/eval, or network operations are present across the scripts. 4. Sanitization: Scoring and direction labels are normalized; however, descriptive fields from input data are processed without specific sanitization.

edge-signal-aggregator