lca-foundry-workflows

Tiangong LCA Skills

Repository: https://github.com/tiangong-lca/skills

Use the skills CLI from https://github.com/vercel-labs/skills to install, update, and manage these skills.

Install the CLI

npm i skills@latest -g

Install

• List available skills (no install):

  npx skills add https://github.com/tiangong-lca/skills --list
  

• Install all skills (project scope by default):

  npx skills add https://github.com/tiangong-lca/skills
  

• Install specific skills:

  npx skills add https://github.com/tiangong-lca/skills --skill flow-hybrid-search --skill process-hybrid-search
  

Target agents and scope

• Target specific agents:

  npx skills add https://github.com/tiangong-lca/skills -a codex -a claude-code
  

• Install globally (user scope):

  npx skills add https://github.com/tiangong-lca/skills -g
  

• Scope notes:
  • Project scope installs into ./<agent>/skills/.
  • Global scope installs into the per-agent user skills directory resolved by the skills CLI on the current platform. Use npx skills list to inspect the exact path on macOS, Linux, or Windows.

Install method

• Interactive installs let you choose:
  • Symlink (recommended)
  • Copy

Update and verify

• List installed skills:

  npx skills list
  

• Check for updates:

  npx skills check
  

• Update all skills:

  npx skills update
  

External runtime skills

This repository owns checked-in TianGong LCA workflow skills. Fast-moving Tiangong KB research skills are consumed from their owning repositories at runtime instead of being mirrored here.

For source-evidence dataset development that needs SCI paper evidence, resolve the latest external skill from tiangong-ai/skills:

npx skills use https://github.com/tiangong-ai/skills --skill tiangong-kb-sci-search --full-depth

Optional local project install:

npx skills add https://github.com/tiangong-ai/skills --skill tiangong-kb-sci-search --agent '*' --yes --full-depth
npx skills update --project --yes

Consuming projects should record the resolved upstream ref and command in task artifacts. Do not copy tiangong-kb-* skill folders into this repository unless the ownership boundary changes deliberately.

Foundry top-level workflows

• $external-dataset-curated-import: BAFU, USLCI, and other structured LCA package imports through CLI conversion, curation queue next/verify, child skills, and publish handoff gates.
• $source-evidence-dataset-development: evidence-driven data creation or update from PDFs, Word files, URLs, APIs, reports, database references, or scientific literature.
• $dataset-rls-maintenance: current-user RLS-scoped cleanup, delete/retire, reference repair, and redo planning for previously imported rows; orchestrates CLI maintenance plans and readback verification without private database access.

Validation

• Validate the canonical CLI-backed wrappers and migration doc guards locally:

  node scripts/validate-skills.mjs
  

• Validate against an unpublished local CLI working tree:

  TIANGONG_LCA_CLI_DIR=/path/to/tiangong-lca-cli \
  node scripts/validate-skills.mjs
  

• Validate only the skills you changed:

  node scripts/validate-skills.mjs lifecycleinventory-qa process-hybrid-search
  

• CI runs the same validation script in .github/workflows/validate-skills.yml after checking out and building tiangong-lca-cli.

Execution note

Skills in this repository are expected to be thin wrappers over the unified tiangong-lca CLI.

Current rules: