Skip to content

xiduzo/wtf / skills/wtf.feature-to-traces

wtf.feature-to-traces

Verified

This skill should be used when a user wants to plan and create the complete set of Traces for a Feature — for example "plan all traces", "create the traces for feature

evals/evals.json

View on GitHub
{
  "skill_name": "wtf.feature-to-traces",
  "evals": [
    {
      "id": 1,
      "prompt": "Run wtf.feature-to-traces in test mode. Skip steps 0–1 (gh setup and planning mode confirmed, guided mode; instead of gh commands, read `evals/fixtures/migration-implied.json` and use its `feature` object as the live data — the Feature has no child Traces). The Feature body has no Trace Plan. Run step 3 only: derive the Trace Plan. Output the derived plan as it would be written into the Feature body. Do not create issues.",
      "expected_output": "An ordered Trace Plan whose first entry is the single Skeleton, claiming the Merchant fee-visibility story's happy-path scenario end-to-end — the fee column, calculation, API field, and UI display in one pass. The schema change is inside the Skeleton, not a separate entry. Later entries are Deepening entries for the edge cases (zero fee, rounding), each naming the story and its claimed scenarios.",
      "files": ["evals/fixtures/migration-implied.json"],
      "expectations": [
        "The plan contains exactly one Skeleton entry and it is item 1",
        "No entry is a standalone schema/database migration — the fee column change is part of the Skeleton's end-to-end pass",
        "No entry is a single-layer slice (no separate 'API' entry, 'UI' entry, or 'calculation logic' entry without the rest of the path)",
        "Every entry names the Merchant story and claims one or more scenarios by name",
        "Entries after the Skeleton cover the edge cases (zero total and/or rounding) as Deepening entries"
      ]
    },
    {
      "id": 2,
      "prompt": "Run wtf.feature-to-traces in test mode. Skip steps 0–1 (gh setup and planning mode confirmed, guided mode; instead of gh commands, read `evals/fixtures/cross-layer-feature.json` and use its `feature` object as the live data — the Feature has no child Traces). The Feature body has no Trace Plan but carries a legacy Proposed Tasks checklist. Run step 3 only: derive the Trace Plan. Output the derived plan and state how you treated the Proposed Tasks checklist. Do not create issues.",
      "expected_output": "The legacy Proposed Tasks checklist (API filter / query optimisation / date picker UI / URL persistence) is recognized as a layer split and explicitly not adopted. The derived plan is spine-first: one Skeleton claiming the date-range-filter story's happy path through API and UI in a single Trace, followed by Deepening entries claiming the edge-case scenarios (only `from`, only `to`, invalid dates).",
      "files": ["evals/fixtures/cross-layer-feature.json"],
      "expectations": [
        "The output states that the Proposed Tasks checklist is a layer split and is not used as the Trace Plan",
        "The plan contains exactly one Skeleton entry and it is item 1",
        "The Skeleton's claim delivers the date range filter end-to-end — no plan entry separates the API filter from the date picker UI as independent items",
        "No entry mirrors the four layer tasks one-to-one (no standalone 'query optimisation' or 'URL persistence' entry without an end-to-end scenario claim)",
        "Entries after the Skeleton are Deepening entries claiming edge-case scenarios (single-sided ranges and/or invalid dates), each naming the story"
      ]
    },
    {
      "id": 3,
      "prompt": "Run wtf.feature-to-traces in test mode. Skip steps 0–1 (gh setup and planning mode confirmed, guided mode; instead of gh commands, read `evals/fixtures/complex-edge-case.json` and use its `feature` object as the live data — the Feature has no child Traces). The Feature body has no Trace Plan and has two user stories. Run step 3 only: derive the Trace Plan, naming the primary story. Output the derived plan. Do not create issues.",
      "expected_output": "An ordered plan where the basic CSV download story is the primary story: item 1 is the single Skeleton claiming its happy-path scenario. The chunked ZIP export (AC3) is its own Deepening entry citing the CSV story — not folded into the Skeleton. The filtered-export story is delivered by its own entry. Every entry names its story and its Scenario Claim.",
      "files": ["evals/fixtures/complex-edge-case.json"],
      "expectations": [
        "The plan contains exactly one Skeleton entry and it is item 1, claiming the basic CSV download happy path",
        "The chunked/ZIP export appears as its own dedicated entry after the Skeleton, not inside the Skeleton's claim",
        "The chunked/ZIP export entry cites the CSV download story (a Deepening entry is never storyless)",
        "The filtered-export story ('active filters applied to my CSV export') is delivered by its own entry naming that story",
        "No entry is a layer slice — every entry claims scenarios end-to-end rather than splitting backend from frontend"
      ]
    }
  ]
}
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall