Profile My skills Saved Collections Edit profile Submit a skill

← all publishers

solanodz

@solanodz source repo

13 published skills

  1. ▌
    Harness E2e · solanodz
    End-to-end and smoke tests that catch what unit tests miss — executable architecture rules and agent-friendly error messages. Use when unit tests pass but the feature fails in runtime or across components.
    0
    installs
  2. ▌
    Harness Audit · solanodz
    Scores your project setup 0–100 and shows what to fix first. Use when the agent keeps failing, before improving a harness, or to see which subsystem is the bottleneck (instructions, state, verification, scope, lifecycle).
    0
    installs
  3. ▌
    Harness Scope · solanodz
    One feature at a time with clear done criteria — less half-finished work. Use when the agent does too much at once, rewrites the task list, or leaves features stuck in progress.
    0
    installs
  4. ▌
    Harness State · solanodz
    Pick up where the last session left off — feature list, progress log, and evidence on disk. Use when context is lost between sessions or long tasks need continuity across days.
    0
    installs
  5. ▌
    Harness Diagnose · solanodz
    When the agent keeps failing, attribute the failure to a harness layer and fix the environment before swapping models. Use for repeated failures, verification gaps, or Lesson 01 diagnostic loops.
    0
    installs
  6. ▌
    Harness Scaffold · solanodz
    Sets up a new project so the agent can start, verify, and resume anytime — AGENTS.md, feature list, progress log, init.sh, handoff. Use when adopting harness engineering or the user asks to set up an agent-ready project.
    0
    installs
  7. ▌
    Harness Evaluator · solanodz
    Separate implementer from verifier — sprint contracts, rubrics, and independent review before accepting work. Use when the agent says done but quality is poor, or for Project 05 self-verification patterns.
    0
    installs
  8. ▌
    Harness Lifecycle · solanodz
    Clean start and handoff every session — init.sh, progress updates, and a restart path. Use when sessions feel chaotic, the next agent can't pick up, or init.sh / handoff files need design.
    0
    installs
  9. ▌
    Harness Environment · solanodz
    Makes the project environment reproducible for agents — toolchain, env vars, services, and init.sh health checks. Use when setup fails, .env is missing, Docker/DB will not start, or a fresh checkout cannot run.
    0
    installs
  10. ▌
    Harness Instructions · solanodz
    Writes a short AGENTS.md the agent actually reads — rules, startup flow, and verification commands. Use when instructions are too long, missing, or the agent keeps asking the same project questions.
    0
    installs
  11. ▌
    Harness Verification · solanodz
    No more "done" without running tests — mandatory checks before marking complete. Use when the agent declares victory too early or init.sh / verification commands need setup.
    0
    installs
  12. ▌
    Harness Observability · solanodz
    See what the agent actually did — captured logs, health checks, and debug commands. Use when the agent says it worked but you can't tell why, or when failures are hard to diagnose.
    0
    installs
  13. ▌
    Harness Repo Knowledge · solanodz
    Builds a short repo map so agents find the right files instead of wandering. Use when the agent explores the whole tree, edits the wrong module, or keeps asking where things live.
    0
    installs