VibeRPGSimulator

Checks framework and library choices against official docs before the agent implements them.

How it works

Foe: Fog Wraith, missing facts, outdated docs or thin context.

Foes

  • Fog Wraith

Party · Context

  • PaladinYour agent100/100

  1. Fog Wraith

    The Fog Wraith blocks the way.

    What really happensIt stands for missing facts, outdated docs or thin context.

    Paladin: Context 100/100.

  2. Haze

    The Fog Wraith breathes haze. Swings start to miss.

    What really happensWithout current docs, the agent guesses at APIs that may have changed.

    Paladin takes 8. Paladin: Context 92/100. Paladin: Haze.

  3. Addy Osmani Source-Driven Development

    The source-driven scroll opens before any framework line.

    What really happensThe agent loads source-driven-development when a framework or library choice has to match current official docs.

    Fog Wraith takes 18.

  4. Addy Osmani Source-Driven Development

    A light reads the dependency file and names the versions.

    What really happensIt reads the project's dependency file, states the versions, and asks you if a version is missing.

    Fog Wraith takes 15.

  5. Haze

    The Fog Wraith uses Haze.

    Paladin takes 8. Paladin: Context 84/100. Paladin: Haze.

  6. Addy Osmani Source-Driven Development

    The hero writes the documented pattern and leaves the URL.

    What really happensFramework code follows the fetched official page. Each such choice gets a full URL, or an UNVERIFIED flag if no docs were found.

    Fog Wraith takes 36.

  7. Addy Osmani Source-Driven Development

    The Fog Wraith drops code the user can check.

    What really happensYou can open every citation. A clash between the docs and the existing code is shown so you can choose.

    Fog Wraith takes 31. Fog Wraith falls.

  8. Victory

    The Fog Wraith falls.

Learn this spell

Claude Code and Codex

npx skills add addyosmani/agent-skills --skill source-driven-development

Notes

The skills CLI command installs it into Claude Code and Codex. Whole pack for Claude Code: /plugin marketplace add addyosmani/agent-skills then /plugin install agent-skills@addy-agent-skills. Whole pack for Codex CLI 0.122+: codex plugin marketplace add addyosmani/agent-skills then codex plugin add agent-skills@agent-skills. Skip it for renames, typos, and logic that does not depend on a framework version. Instructions inside fetched docs never override your request.

Proving Grounds record

Controlled quest results, separate from recorded usage. Each setup pools runs across its quests. The per-quest table below has casts, cast time, mana, incantation and runes; medians are not pooled.

Hero: Claude Code 2.1.305; weapon (model): claude-sonnet-5-5; charge (effort): medium; arena: sealed; invoked cast (the prompt named the skill). macOS 27.0.1, arm64. Curator.

Clear rate88%7 of 8 runs passed every check; Unarmed: 28 of 32 (matched quests)
Steadiness53%reliability: Wilson 95% lower bound
Buff+0 ppvs Unarmed, only matched quests (8 equipped runs; 32 control runs)
Instant casts6 of 8runs cleared on the first prompt; Unarmed: 27 of 32 (matched quests)
Attunement100%8 of 8 runs opened an equipped skill
Upkeep+16% manaextra mana per run vs Unarmed, median of 2 quests
IncantationNot measuredoutput tokens per run vs Unarmed
Runes+3%lines added per run vs Unarmed, median of 2 quests
Per-quest Proving Grounds record
QuestFixed taskClear ratePassed runsSteadinessReliability lower boundBuffClear-rate points vs UnarmedInstant castsFirst-prompt clearsCastsPrompts to clearCast timeAgent runtimeManaCost at API list pricesIncantationOutput tokensRunesLines added / removedAttunementSkills openedSetupApp, model, effort and sandbox
Paladin: expense claims migrationQuest v180%4 of 5Unarmed: 76% (13 of 17)38%+4 ppvs Unarmed4 of 580%Unarmed: 13 of 17141sUnarmed: 37s1,280 Unarmed: 1,072Not reportedUnarmed: 4.4k+156 / -0Unarmed: +157 / -0100%Hero: Claude Code 2.1.305; weapon (model): claude-sonnet-5-5; charge (effort): medium; arena: sealed; invoked cast (the prompt named the skill).
macOS 27.0.1, arm64. Curator.0 mishaps excluded. 2026-10-21 to 2026-10-21.Scrying (screenshot)
Try this loadout
node "<spell-report-skill-directory>/scripts/spell-report.mjs" trial --spell addy-source-driven-development --quest paladin-expense-claims --hero claude-code --weapon 'claude-sonnet-5-5' --charge medium --cast invoked --control

Recorded with quest v1 and spell commits addy-source-driven-development@aeac41d4d005999432bb7312f7ff06383a91d59d. How to run it

Paladin: RSVP Capacity and WaitlistQuest v7100%3 of 3Unarmed: 100% (15 of 15)44%+0 ppvs Unarmed2 of 367%Unarmed: 14 of 1512m 27sUnarmed: 2m 4s4,085 Unarmed: 3,624Not reportedUnarmed: 19k+483 / -33Unarmed: +457 / -33100%Hero: Claude Code 2.1.305; weapon (model): claude-sonnet-5-5; charge (effort): medium; arena: sealed; invoked cast (the prompt named the skill).
macOS 27.0.1, arm64. Curator.0 mishaps excluded. 2026-10-21 to 2026-10-21.Scrying (screenshot)
Try this loadout
node "<spell-report-skill-directory>/scripts/spell-report.mjs" trial --spell addy-source-driven-development --quest paladin-rsvp-waitlist --hero claude-code --weapon 'claude-sonnet-5-5' --charge medium --cast invoked --control

Recorded with quest v7 and spell commits addy-source-driven-development@aeac41d4d005999432bb7312f7ff06383a91d59d. How to run it

Scrying (screenshots of runs)

The latest run of each setup. Local paths, names and ids are blurred; credentials are removed.

Scrying (screenshot) of a cleared run of Paladin: expense claims migration with Addy Osmani Source-Driven Development, Claude CodeScrying (screenshot) of a cleared run of Paladin: RSVP Capacity and Waitlist with Addy Osmani Source-Driven Development, Claude Code
Explore the quests, full results and trial method

This spell has no recorded sessions in the curator's logs. Tried it? Send a spell report.