Recipe Fit Assessor

recipe-fit-assessor

Purpose

Map a user's stated goal to the closest documented recipe or pack (product-hero-v0, golden paths, workflow docs) and surface mismatches before expensive planning or runtime work.

Quick start

• capture goal + constraints + deliverables
• compare against catalog rows in references/recipe-catalog.md
• assign fit verdict (Strong / Partial / Poor)
• list gap fixes or recommend a different workflow entrypoint

When to use

• user asks "which recipe / pack fits my idea"
• scope drifts between product hero, portrait, interior, web GLB, batch renders
• before locking intent-to-3d-brief-writer output to a pack

When not to use

• runtime execution or MCP troubleshooting (use runtime skills)
• pure creative brainstorming without a delivery shape (use blender-brainstorming first)
• legal/license checks (use asset-license-checker)

Trigger phrases

• "what recipe should I use"
• "does product hero fit"
• "closest workflow for X"
• "which pack"

Prerequisites / readiness

• rough goal statement exists
• intended outputs known (still image, GLB, animation, web, print)
• audience skill level known (non-Blender user assumed default)

Input schema

Required inputs

| Input | Why it is required | |---|---| | Goal summary | Defines fit target | | Deliverable type | Maps to recipe outputs | | Constraints (time, realism, web vs cinematic) | Drives fit verdict |

Optional inputs

| Input | Use | |---|---| | Reference links / moodboards | Clarify genre fit | | Prior pack choice | Re-evaluate fit |

Assumptions to confirm

• User accepts switching packs if fit is Poor.
• Recipes under docs/recipes/ and packs under packs/ are the authoritative catalog.

Output schema

Primary output

• recommended recipe / pack name(s)
• fit verdict + rationale
• gap list (what must change for Strong fit)
• suggested next skill (intent-to-3d-brief-writer, domain planner, etc.)

Evidence / caveat output

Runtime status: Not Run | Attempted | Produced | Verified | Failed | Blocked / Not Run
Artifact status: Not Run | Not Produced | Produced | Verified | Failed
Evidence used: 
Limitations: 

Required laws

• ../../laws/evidence-before-done.md
• ../../laws/non-blender-user-language.md
• ../../laws/no-arbitrary-python-interface.md
• ../../laws/official-runtime-only.md

Official runtime boundary

Planning-only skill. Does not configure MCP, Blender, or hosts. When mentioning runtime, use the 2-path + CLI appendix model from docs/runtime-stack-strategy.md only as context for whether a workflow expects interactive MCP vs batch CLI — never claim runtime success.

Operating procedure

1. Normalize goal using references/scope-gap-rules.md.
2. Score against each catalog row in references/recipe-catalog.md.
3. Pick Strong fit if any; else best Partial with explicit gaps.
4. Emit recommendation block from references/workflow-recommendation-template.md.
5. Hand off to brief writer or narrow planner per fit.

Mode handling

Text-only mode

Fit from stated intent only; flag unknowns as gaps.

Runtime-ready mode

Same as text-only; optionally note if user demands CLI-only automation.

Blocked runtime mode

Still usable; note runtime unavailable does not change recipe fit.

Validation checklist

• [ ] Catalog consulted
• [ ] Verdict assigned
• [ ] Gaps listed or "none"
• [ ] Next skill named
• [ ] No false runtime claims

Pass / Warn / Fail rubric

| Verdict | Criteria | |---|---| | Pass | Strong fit with documented recipe/pack + clear next skill | | Warn | Partial fit; user must accept gaps or expand scope | | Fail | Poor fit but user insists on wrong pack without mitigations |

Cross-skill handoff

• From brainstorming → ../blender-brainstorming/SKILL.md
• To structured brief → ../intent-to-3d-brief-writer/SKILL.md
• Product hero path → ../product-hero-scene-planner/SKILL.md
• Help routing → ../blendops-help/SKILL.md

References

• references/recipe-catalog.md
• references/scope-gap-rules.md
• references/workflow-recommendation-template.md
• ../../docs/recipes/README.md
• ../../packs/product-hero-v0/PACK.md

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: ThanhNguyxnOrg
• Source: ThanhNguyxnOrg/blendops
• License: MIT
• Homepage: https://github.com/ThanhNguyxnOrg/blendops

Install and usage instructions live in the source repository linked above.