The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple local shell scripts (phase1_discovery.sh and phase2_analysis.sh) to perform project scanning.
Evidence: bash ~/.gemini/antigravity/skills/proje-analizcisi/scripts/phase1_discovery.sh and bash ~/.gemini/antigravity/skills/proje-analizcisi/scripts/phase2_analysis.sh.
The scripts utilize various CLI tools including find, grep, awk, git, and xargs to process project data.
[DATA_EXFILTRATION]: The skill is designed to search for and extract sensitive information such as hardcoded credentials and environment variables for reporting purposes.
Evidence: Grep patterns for password, api_key, and .env files in scripts/phase2_analysis.sh.
Accessing these values brings them into the agent's context, though no external transmission (exfiltration) to third-party domains was detected in the scripts.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it processes untrusted data from project files.
Ingestion points: Reads files throughout the analyzed project directory using find, cat, and grep in phase1_discovery.sh and phase2_analysis.sh.
Boundary markers: None identified in SKILL.md to prevent the agent from obeying instructions embedded within the code being analyzed.
Capability inventory: The skill can execute bash scripts and create/write files (mkdir -p docs/analyze).
Sanitization: No evidence of sanitization or escaping of external content before it is processed for report generation.
[REMOTE_CODE_EXECUTION]: The skill uses python3 -c to execute dynamic Python strings for JSON parsing of package.json files.
Evidence: Hardcoded Python logic in scripts/phase1_discovery.sh and scripts/phase2_analysis.sh used to check dependencies.

proje-analizcisi