Profile My skills Saved Collections Edit profile Submit a skill

← all publishers

LSTR-1

@lstr-1 source repo

4 published skills

  1. ▌
    Self Test · lstr-1
    Self-evaluates the agent's output under controlled fault injection using deterministic rule-based graders. Invoke to verify a deliverable before delivery, to test robustness against environmental degradation, or when acceptance must not rely on LLM-as-judge. Trigger phrases "test this", "verify my output", "check robustness", "self-test", or as the Verify stage of the metacognitive loop.
    0
    installs
  2. ▌
    Self Evolve · lstr-1
    Reflects on a completed task's outcome and distills reusable atomic rules into persistent state (skills and memory), governed by the self-evolution policy. Invoke after a task passes its deterministic grader, on "/reflect", or as the Evolve stage of the metacognitive loop. Trigger phrases "reflect on this task", "/reflect", "learn from this", "distill a rule".
    0
    installs
  3. ▌
    Self Discover · lstr-1
    Builds a zero-shot runtime capability inventory for the LSTR Solo Harness. Invoke when the agent needs to discover its own capabilities — trigger phrases "discover capabilities", "what can you do", "inventory your tools", "/discover", or at SessionStart bootstrap. Produces a capability table (capability, source, trigger condition, confidence) from skills, MCP tools, rules, and memory.
    0
    installs
  4. ▌
    Metacognitive Plan · lstr-1
    Orchestrates the LSTR 6-step metacognitive reasoning chain and produces a decision-complete plan against verifiable constraints. Invoke for architecture decisions, multi-constraint trade-offs, incident triage, or any task requiring auditable reasoning. Trigger phrases "plan this", "/plan" (delegating to the reasoning chain), "apply the LSTR framework", "reason through this", or any query where constraints compete and a plan is needed.
    0
    installs