AGENT SKILL

simready-foundation-update-feature-adapter

Use for updating SimReady feature adapters, USD mutation logic, tests, and source/target metadata.

License 许可
Apache-2.0
所属仓库
leemiracle/work4ai
最近更新
2026-11-12
Stars
★ 4
Forks
0

01 前言 Front Matter

name
simready-foundation-update-feature-adapter
license
Apache-2.0
metadata
{ "tags": [ "simready", "adapter", "maintenance" ], "author": "Shaad Boochoon <sboochoon@nvidia.com>" }
description
Use for updating SimReady feature adapters, USD mutation logic, tests, and source/target metadata.

03 SKILL.md 正文

SimReady Update Feature Adapter

Purpose

Use this skill when an existing feature adapter no longer matches exact source/target FET_###_<RUNTIME> feature contracts, produces invalid USD, misses a required mutation, or needs new profile-version support.

Prerequisites

Before editing, read:

  • AGENTS.md
  • nv_core/sr_specs/docs/guides/feature_adapters/feature_adapters.md
  • target adapter file
  • source and target feature manifests
  • source and target profile versions
  • existing adapter tests or validation reports

Inputs

Collect or infer:

Input Requirement
adapter_path Existing adapter file.
source_feature Input feature name / manifest id and integer version currently supported.
target_feature Output feature name / manifest id and integer version currently supported.
bug_or_gap Missing mutation, bad metadata, invalid output, non-idempotence, crash, or profile drift.
new_feature_versions Updated integer feature versions if the adapter should support a new path.
test_assets Assets or reports that show the problem.

Instructions

Use this checklist when changing the repository:

  1. Compare adapter metadata with actual source/target manifests and profile versions. Confirm feature names use FET_###_<RUNTIME> and feature versions are exact integers.
  2. Reproduce or inspect the failed upgrade when possible.
  3. Determine whether to update the existing adapter or add a new adapter for a new exact version path. Prefer adding a new path when published source/target versions must keep their old behavior.
  4. Patch transformation logic narrowly:
    • preserve unrelated data
    • keep output deterministic
    • avoid source-stage mutation
    • save output stage correctly
  5. Update metadata only when the supported feature path changes.
  6. Update tests or manual validation instructions to cover the corrected path.
  7. Update profile migration notes if adapter behavior affects profile upgrades.
  8. Confirm every changed feature difference in the profile transformation still has a direct adapter path or documented blocker.
  9. Validate source-to-target behavior against the relevant feature/profile requirements when tooling is available.

Examples

Example request:

Update a SimReady feature adapter for a new exact source and target feature version.

Expected result summary:

changed_files: updated docs, manifests, indexes, validators, or adapters
validation: focused checks for the changed contract
remaining_gaps: downstream feature, profile, adapter, or runtime-test follow-up

Policies

  • Do not change an adapter to target a different feature version without preserving the old upgrade path when it is still published.
  • Prefer adding a new adapter for new feature versions over mutating old behavior.
  • Do not add broad catch-all mutation logic that could damage assets outside the adapter contract.
  • If a required mutation needs user intent, report a blocked adapter path.

Limitations

  • Do not silently change published contracts; add versions when behavior changes.
  • Do not update docs without checking validator, manifest, profile, adapter, and runtime-test impact.
  • Do not broaden a change beyond the selected requirement, capability, feature, profile, validator, or adapter.

Troubleshooting

  • Error: the requested change alters a published contract. Solution: create a new version and preserve the old artifact.
  • Error: docs and machine-readable manifests diverge. Solution: make the JSON/TOML and markdown agree, then rerun focused checks.
  • Error: downstream profile or adapter impact is unclear. Solution: list the affected artifacts before making broad edits.

Resources

  • assets/openai.yaml preserves optional UI metadata for clients that read skill display hints. It is not required for the workflow.

Summary Format

Report:

Field Meaning
adapter_path Adapter changed.
feature_path Input and output feature names and integer versions.
change_type Bug fix, new version path, or docs/test update.
mutation_summary USD behavior changed.
tests Tests/assets run.
profile_followup Profile docs or migration notes changed/needed.
validation Remaining gaps.

04 同仓库更多 Skills

leemiracle
leemiracle/work4ai

agentic-ecology-ui

Guidelines and templates for creating a generic web UI to visualize, rank, filter, and annotate vector databases. The UI has a hierarchical structure covering database selection, label selection/definition, ranking/filtering configuration, and row visualization/annotation. Use when building, serving, or adapting an interactive web UI for Hoplite vector databases.

★ 4
leemiracle
leemiracle/work4ai

agentic-ecology-init

Initializes a local uv-managed project directory for agentic ecology workloads. Sets up Python dependencies using reference pyproject.toml and uv.lock, configures workspace rules, and ensures Agentic Ecology skills are discoverable. Use when initializing a new project workspace or bootstrapping a clean environment.

★ 4
leemiracle
leemiracle/work4ai

agentic-ecology-camera-traps

Provides capabilities to run SpeciesNet detector and classifier on camera trap images, extract crop-level feature embeddings, and populate a Hoplite vector database for downstream search and agile modeling. Use when ingesting camera trap images, classifying wildlife photos, or indexing camera trap data.

★ 4
leemiracle
leemiracle/work4ai

agentic-ecology-bioacoustics

Provides bioacoustic analysis capabilities for ecologists and researchers using the perch-hoplite Python package. A typical use case is to use agile modeling to bootstrap the creation and deployment of a bespoke detector for targeted species on an existing collection of passive acoustic monitoring recordings. Use when processing and annotating audio recordings for bioacoustic applications.

★ 4
leemiracle
leemiracle/work4ai

sample-skill

Sample skill used by agent-env integration tests. Exercises SKILL.md parsing plus supporting scripts and references.

★ 4
leemiracle
leemiracle/work4ai

simready-foundation-validate-foundation-change

Use for auditing SimReady requirement, validator, feature, profile, adapter, test, and skill consistency.

★ 4