Panda Suite Docs
v1.9.2GitHub ↗
Panda Suite / Upstream maintenance

Upstream integration

All thirty tracked upstream files are included in this checkout. No recursive clone, network bootstrap, symlink traversal, or runtime fetch is needed to use the twenty packaged skills in the working checkout.

upstream-lock.json maps upstream source paths to local destinations and records the repository URL, full commit SHA, license state, and per-file SHA-256 hash. Primary skill files live in the discoverable skills/<name>/ locations. The exact pinned Panda Debugger SKILL.md is preserved at vendor/panda-debugger/SKILL.md; the active skills/panda-debugger/SKILL.md includes Suite-specific success guidance. Other upstream tests, READMEs, and old manifests live under the primary skill folders; archived manifests are not loaded by the Suite plugin.

Verify and recover

python3 scripts/vendor-upstreams.py
python3 scripts/vendor-upstreams.py --restore

The first command is offline. The second fetches pinned commits only if files are missing, verifies their hashes before writing, and refuses existing files with local modifications. It restores missing files; it is not a reset command. No untrusted archive is extracted and no upstream script is executed.

Updating a pin

  1. Inspect the new upstream revision, references, tests, license, and manifests in a separate clone. Do not edit the original GitHub repository.
  2. Verify the existing lock first. If local modifications exist, review and preserve them before importing new source.
  3. Copy the selected revision's tracked files using the same source/destination mapping. Keep the pristine Debugger SKILL.md at vendor/panda-debugger/SKILL.md and maintain the active Suite instructions at skills/panda-debugger/SKILL.md; Buddhist Method's SKILL.md, references/*, and LICENSE go under skills/buddhist-method/; other tracked files go under the corresponding skill directory. Store original dot-directory manifests under source-metadata/ so they remain provenance rather than active package metadata.
  4. Update the full commit SHA, complete file list, and SHA-256 values in the lock. Review additions and removals explicitly. Keep Suite-authored changes out of the upstream snapshots; retain all attribution and license notices.
  5. Update provenance notes and run structural, drift, host, and behavioral checks. A newly changed upstream workflow may invalidate earlier evaluation results.

Independent installations are also possible. Install the original projects using their own instructions and install Suite specialists individually. Avoid installing two copies with the same skill name: hosts may expose duplicates rather than merge them. Lead can select an available original skill by name; it does not need a wrapper or a bundled replacement.

Panda Debugger's absent upstream license is recorded in source notices. The Suite license does not grant rights to those upstream files. Public distribution terms remain unresolved.