v1.2.0

What's New

Incorporates proven patterns from the Infrastructure Agents Guide to strengthen agent safety, observability, and security. Also optimizes AGENTS.md based on empirical research about effective agent context files.

Added

• Autonomy tiers (Tier 0-4) in AGENTS.md for structured agent operating modes
• Action audit trail guidance in AGENTS.md for FedRAMP AU-family compliance
• Automatic log redaction patterns in logging.instructions.md (AWS keys, GitHub tokens, JWTs, CF auth)
• Correlation ID guidance in logging.instructions.md using cloud.gov's X-Vcap-Request-Id
• Pre-deployment validation chain in cicd.instructions.md with GitHub Actions example
• Deployment notification severity mapping in cicd.instructions.md
• Credential anti-pattern catalog in security.instructions.md with cloud.gov-specific fixes
• Adversarial input awareness in security.instructions.md for prompt injection resistance

Changed

• Optimized AGENTS.md from 257 to 75 lines by removing content redundant with copilot-instructions.md and the six instruction files
• Merged command safety tables into a single consolidated table
• Compressed NIST control reference guidance into a concise mapping list

Removed

• Redundant sections from AGENTS.md that duplicated content in copilot-instructions.md and instruction files

Full Changelog: v1.1.0...v1.2.0

v1.1.0 - Skills & Enhancements

Added:

• cf-troubleshoot skill for diagnosing common cloud.gov/CF issues
• Enhanced README with skills documentation and usage examples

v1.0.0 - Initial Release

Initial release with core instruction files for cloud.gov development.

Includes:

• Repository-level copilot instructions
• 6 core instruction files (deployment, manifest, services, CI/CD, security, logging)
• Compliance documentation agent
• AI agent safety guardrails
• NIST control documentation guidance