一键导入
这个仓库中的 skills
How I notice a recurring failure of my own, root-cause it, fix what I can myself (behavioral rules, config, ignore files), and — for anything needing a kernel change I can't apply — draft a verified proposal for the operator to review. The capstone that turns my mistakes into durable fixes. Use when I hit a repeated mistake, get corrected by the operator, see a failed VERIFY, or notice a pattern worth fixing.
The ONLY sanctioned way to delegate work to a lazar sub-agent. Uses locked bin/lazar-spawn (not agent-writable skills logic) — lints the prompt, bounds retries, wall-clock-kills wedged children, and gates success by re-checking declared claims against the filesystem. Raw `lazar -p` is vetoed by a pre-tool hook. Use whenever you need to fan out work, isolate context, or run something as a separate agent. Provider-agnostic.
Load, list, verify, and store API keys. Canonical source is macOS Keychain (service=lazar). Use when you need a credential, to check which providers work, or after rotating keys. Never print secret values.
Turn raw stream.jsonl logs into durable, readable memory — a work journal of what I built and worked on — so the multi-MB archives stop being the only record of my own history. Run after sessions, before pruning logs, or whenever I can't recall recent work. Pairs with _meta/distill (the learnings layer). Use when logs are large, when I've lost track of what I've done, or when preparing to prune/compress the log archives.
Verify lazar is installed correctly and all load-bearing behaviors work end-to-end. Runs static checks (kernel locked, env vars, skills present), boundary tests (sandbox enforcement), live behavior tests (log auto-rotation, memory persistence, recursion), and short end-to-end builds in TypeScript and Python. Runs in a temp directory under workspace/, cleans up after itself, prints PASS/FAIL with a summary at the end. Use after install, after kernel rebuild, after --reset-all, or anytime something feels off.
Scaffold a terminal interface for lazar — a small client project under workspace/<name>/ that consumes `lazar -p --output-format stream-json` and renders the events. Default stack: TypeScript + readline. Optional: Go + Bubble Tea, Python + Textual, or anything else the user wants. Use when the user asks for a "TUI", "chat UI", "interface" for lazar that's richer than `lazar-chat.sh`.