engram

engram

spectra-g

Uses Git history and organizational memory to help AI agents avoid breaking changes by identifying hidden dependencies and co-committed files.

597 views2Local (stdio)

What it does

  • Analyze blast radius of file changes using Git history
  • Save project context and architectural decisions
  • Read historical notes about code areas
  • Track usage metrics and risk distributions

Best for

AI agents making code changesPreventing regressions in large codebasesPreserving institutional knowledge across conversations
Local-first with zero telemetryDetects hidden dependencies without code imports

Tools (4)

get_impact_analysis

Analyzes the blast radius and coupling of files to prevent breaking changes. This tool reveals which other files are frequently co-committed with the target file, helping you understand what else might be affected by your changes. Use cases: bug fixes, feature additions, refactoring, code review.

save_project_note

Save important context, decisions, or learnings about a file or symbol for future reference. Use this when you discover important information about code behavior, architectural decisions, bug fixes, or gotchas that would be valuable for future work. Notes are stored persistently and will appear in future impact analysis results.

read_project_notes

Read project notes (memories) about files before working on them. Use this when starting work on a file or investigating an area of the codebase to see if there are important notes, learnings, or context that have been saved. Can search by query text, filter by file path, or list all notes.

get_usage_metrics

Get usage statistics for this repository including how many analyses have been performed, risk distributions, notes created, and performance metrics. Useful for understanding your usage patterns and the health of the codebase analysis.

Alternatives