context-compaction
Context Compaction
Preserve essential state when context window fills up.
Always Preserve
| Item | Why |
|---|---|
| Test names + output (pass/fail) | Regression detection, reproducibility |
| Build status (success/error + msg) | Know if in broken state |
| Files modified (path + why) | Track changes, enable rollback |
| Open TODOs with state | Prevents losing WIP |
| Env vars set this session | Re-establish environment |
| Decisions made (WHY) | Enables contextual future decisions |
DO NOT Compress
- Exact file paths
- Error messages (verbatim)
- Test names
- Numeric results (counts, sizes, checksums)
Example Output
Tests: 2 fail (test_embed_vector, test_turso_insert)
Build: error - "missing trait impl Write for &str"
Files: do-memory-core/src/embed.rs (added async embed)
TODOs: [in_progress] refactor embed.rs - 60% done
Env: TURSO_DATABASE_URL=libsql://...
Decision: switched to async embed to fix blocking in hot path
See Also
- Session State Preservation - Full guidelines
- Token Efficiency - Tool selection priority
More from d-o-hub/rust-self-learning-memory
loop-agent
Execute workflow agents iteratively for refinement and progressive improvement until quality criteria are met. Use when tasks require repetitive refinement, multi-iteration improvements, progressive optimization, or feedback loops until convergence.
51web-search-researcher
Research topics using web search and content fetching to find accurate, current information. Use when you need modern information, official documentation, best practices, technical solutions, or comparisons beyond your training data.
46perplexity-researcher-reasoning-pro
Highest level of research and reasoning capabilities for complex decision-making with significant consequences, strategic planning, technical architecture decisions, multi-stakeholder problems, or high-complexity troubleshooting requiring expert-level judgment and sophisticated reasoning chains. Prioritizes actively maintained repositories and validates website sources for 2025 relevance.
44context-retrieval
Retrieve relevant episodic context from memory for informed decision-making. Use when you need past episodes, patterns, or solutions to similar tasks.
44rust-code-quality
Perform comprehensive Rust code quality reviews against best practices for async Rust, error handling, testing, and project structure
43codebase-analyzer
Analyze implementation details, trace data flow, explain technical workings, locate files, and consolidate codebases. Use when you need to understand HOW code works, find file locations, or assess technical debt.
40