release: v0.16.1

[bug-ops]

Summary

• Bump version from 0.16.0 to 0.16.1
• Finalize CHANGELOG.md [Unreleased] section into [0.16.1] - 2026-03-21
• Update tests badge to 6293
• Update splash screen snapshot for new version
• Merge comprehensive documentation update from docs: comprehensive documentation update for post-v0.16.0 features #2085

Changelog highlights

Fixed

• fix(memory): MAGMA entity type prompt/enum mismatch (fix(memory): MAGMA entity type prompt/enum mismatch — "technology" not recognized #2079)
• fix(sanitizer): memory_search false positive injection flags (fix(sanitizer): memory_search tool output path not covered by MemorySourceHint fix (#2053) #2057)
• fix(core): compaction probe rejection UX (fix(compact): /compact shows 'success' message when probe HardFail rejects compaction #2058)
• fix(scheduler): task injection format (fix(scheduler): task injection format is ambiguous — LLM cancels bash tasks instead of executing them #2073)
• fix(llm): OpenAI content null serialization (bug(llm): OpenAI API 400 Bad Request on skill documentation #1952)
• fix(channels): AnyChannel trait method forwarding, confirm timeouts

Added

• feat(memory): SYNAPSE spreading activation retrieval (research(memory): SYNAPSE spreading activation retrieval over entity graph #1888)
• feat(memory): MAGMA multi-graph typed edges (research(memory): MAGMA multi-graph memory with causal/temporal/semantic/entity edge types #1821)
• feat(context): HiAgent subgoal-aware compaction (research(context): HiAgent subgoal-aware context compaction for long-horizon task coherence #2022)
• feat(memory): compaction probe validation (research: task-continuation metric for post-compaction validation #1609)
• feat(memory): semantic response caching (Research: semantic response caching for LLM API cost reduction #1521)
• feat(memory): write-time importance scoring (research(memory): write-time importance scoring for improved memory retrieval #2021)
• feat(tools): dynamic tool schema filtering (research(tools): dynamic tool schema filtering to reduce context waste and improve selection accuracy #2020)
• feat(tools): tool dependency graph (research(tools): tool dependency graph for sequential tool availability #2024)
• feat(core): multi-language FeedbackDetector (7 languages, FeedbackDetector only supports English language patterns #1424)
• feat(orchestration): plan template caching (research(orchestration): agentic plan caching for LLM planner cost reduction (APC) #1856)

Checklist

• Version updated in workspace manifest (all 20 crates inherit)
• Cargo.lock regenerated
• CHANGELOG.md has [0.16.1] section with date
• README tests badge updated to 6293
• Splash screen snapshot updated
• Documentation merged from docs: comprehensive documentation update for post-v0.16.0 features #2085
• All 6293 tests pass
• cargo fmt + clippy clean

After merge

git tag v0.16.1 && git push origin v0.16.1