chore: scrub internal identifiers from tests, fixtures and comments

[polakamtejas]

Summary

Removes identifying internal data from tests, fixtures and comments. Behaviour does not change.

• Tests and fixtures: internal eval names, run and instance ids, task and delivery layouts, the owning team name and internal image names are now neutral placeholders (multi-env, example-team, example/mcp-server:v1, /app/workspace, …). The golden wire-format fixtures and the tests that pin them change together.
• Comments and docstrings: removed or reworded where they named internal repos, pipelines or types, cited internal PR and review numbers, or quoted production measurements. Where a comment explained something, the rationale stays and only the internal detail goes.

Deliberately not touched:

• Model names in tests and the UI model list, pending a separate decision.
• The private-server gateway integration tests and their docs/CI references.
• Areas the open PRs already cover: refactor(attribution)!: remove project_id and the LiteLLM cost attribution built on it #45 (project_id / LiteLLM attribution, which also removes the attribution comment in deploy_agent.py), refactor(artifact)!: store a file universe's bundle as bundle_object_url #44 (bundle_s3_url), chore(explorer): clean up leftovers in the docs page, runner and API #41 (docs page and runner).

Test plan

• pytest tst/unit packages/agentenv-protocol/tests: 5783 passed, 13 skipped. The 2 failures in test_mcp_cli_builder.py are environmental and also happen on main: the generated CLI's interpreter has no click.
• Golden wire-format tests pass with the updated fixtures.
• check_plugin_api.py: no break to the plugin surface.
• Explorer UI: tsc --noEmit, npm run test:smoke, npm run build:static.

🤖 Generated with Claude Code

Confidence Score: 5/5

The PR appears safe to merge; no new actionable issue or outstanding previous finding was established.

What we checked:

• Short file uploads: The reader checks the total bytes written against the file size and raises instead of returning the partial file.

Summary

This PR replaces identifying examples in tests and golden fixtures with neutral placeholders, and removes internal references from comments. The test scenarios and runtime behavior stay the same.

• Test inputs now use generic team names, environment IDs, image names, and artifact paths.
• Golden JSON fixtures change alongside the test data that produces them.
• Comments keep their technical explanations without internal project names, run IDs, issue numbers, or production details.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart LR
  Tag[Version tag] --> Protocol[Publish protocol]
  Protocol --> Framework[Publish framework]
  Framework --> Release[Create GitHub release]
  CUA[CUA file] --> Size[Check file size]
  Size --> Small[Fetch small file]
  Size --> Large[Read large file in chunks]
  Small --> Upload[Upload artifact]
  Large --> Upload

Loading

_{Reviews (2) · Last reviewed commit: "Merge branch 'main' into tejas/scrub-int..."}