Skills Reference

31 slash-command skills

Skills are slash commands that expand into specialized prompts. They are namespaced by plugin, so each is invoked as /chitta:<name>. They guide Claude through structured workflows — from soul initialization to autonomous development rituals.

Status as of 2026-09-20 · Names follow the repository's skill directories; maintenance suffixes were retained during the rename.

Memory Operations

8 skills ▾
/chitta:initInitialize the soul with foundational beliefs and wisdom
/chitta:checkpointSave task status, decisions, active files, and next steps as a checkpoint
/chitta:reawakenRestore task context from a checkpoint using Pratyabhijñā (recognition)
/chitta:exploreRLM71-style recursive memory exploration — navigate the memory graph instead of taking a top-k dump
/chitta:healthSoul system health check with remediation
/chitta:introspectSoul self-examination (Svādhyāya)
/chitta:migrateImport soul data from SQLite or shared chitta files
/chitta:rememberSave a memory to chitta

Session Continuity

3 skills ▾
/chitta:recapRebuild working context from transcripts and chitta memories in approximately 1500 tokens
/chitta:resumeResume a thread by loading its ~800-token context capsule
/chitta:tasksBrowse and resume tasks, threads, and background jobs across sessions

Reasoning

2 skills ▾
/chitta:antahkaranaMulti-perspective reasoning through cognitive voices
/chitta:ultrathinkAnalyze assumptions, constraints, and alternative explanations

Development

6 skills ▾
/chitta:codebase-learnLearn codebase structure with tree-sitter parsing and SSL pattern extraction
/chitta:yajñaAutonomous development ritual with role-based coordination. Loops until complete using specialized agents: hotṛ→research, adhvaryu→implement, udgātṛ→test
/chitta:long-taskInitialize or resume a long-running task session across multiple sessions
/chitta:shepherdAutonomous pipeline monitor using a sense-think-act loop. Watches snakemake and nextflow jobs, detects errors, applies fixes from memory, restarts on failure
/chitta:kriyaReview soul discoveries one by one, accept or discard each, implement the accepted ones, build chitta, and optionally release
/chitta:prog-reviewFirst-principles review — question requirements, delete unnecessary parts, simplify, optimize with evidence, automate last

Learning & Curiosity

3 skills ▾
/chitta:learnLearn from web sources, generate a SKILL.md, and store key patterns in soul memory
/chitta:dreamTrigger autonomous curiosity-driven exploration. The soul picks a topic from memory gaps or curiosity seeds, searches the web, and stores what it finds as dream-tagged memories
/chitta:hint-corpusFine-tune the Qwen3-0.6B hint model — corpus generation, LoRA/unsloth, GGUF export, Ollama

Maintenance

3 skills ▾
/chitta:epsilon-yajnaConvert stored memories to SSL v0.4 format and check that the result can be decoded
/chitta:mermaidRender Mermaid diagrams as ASCII/Unicode art or SVG
/chitta:distill-pendingProcess pending transcript distillation

System

5 skills ▾
/chitta:cc-soul-setupBuild from source; skill suffix retained after rename (C++20 compiler, CMake, make).
/chitta:cc-soul-updateUpdate binaries; skill suffix retained after rename.
/chitta:cc-soul-daemonManage the daemon; skill suffix retained after rename.
/chitta:cc-soul-shutdownGraceful shutdown; skill suffix retained after rename.
/chitta:cc-soul-mcpConfigure MCP; skill suffix retained after rename (mcp__chitta__*).

Benchmarking

1 skill ▾
/chitta:locomo-benchmarkRun the LoCoMo17 benchmark for long-term conversational memory
← Back to chitta

References

  1. Adyasha Maharana, Dong-Ho Lee, Sergey Tulyakov, Mohit Bansal, Francesco Barbieri, and Yuwei Fang. Evaluating Very Long-Term Conversational Memory of LLM Agents. ACL (2024); arXiv:2402.17753. https://arxiv.org/abs/2402.17753 https://aclanthology.org/2024.acl-long.747/
  2. Alex L. Zhang, Tim Kraska, and Omar Khattab. Recursive Language Models. arXiv:2512.24601 (2025; revised 2026). https://arxiv.org/abs/2512.24601