fix(memory): track injected skill experiences

[fukalous]

Description

The trajectory lineage pipeline only credited successful explicit read/multi_read calls. The Claude Code skill-experience hook injects retrieved Experience digests as text instead, so trajectories created from those sessions never carried source-experience tags and application statistics remained empty.

This change recognizes the exact envelope and bullet format emitted by the bundled hook, then sends each candidate through the existing current-user Experience URI validation before adding it to lineage. Explicit reads and injected URIs share the existing ordered de-duplication path.

Human Involvement

• A human participated in the implementation or review loop
• This PR was generated entirely by AI agents without human participation in the loop

Related Issue

Fixes #5405

Type of Change

• Bug fix (non-breaking change that fixes an issue)
• New feature (non-breaking change that adds functionality)
• Breaking change (fix or feature that causes existing functionality to not work as expected)
• Documentation update
• Refactoring (no functional changes)
• Test update

Changes Made

• Parse Experience URIs only inside the canonical skill-experience / experience-digest context envelope.
• Reuse canonical_experience_uri() to reject other-user and non-Experience paths.
• Preserve message order and deduplicate an injected URI that is also explicitly read.

Testing

• I have added tests that prove my fix is effective or that my feature works
• New and existing relevant unit tests pass locally with my changes
• I have tested this on the following platforms:
  • Linux
  • macOS
  • Windows

Commands run:

.venv/bin/python -m pytest -q --no-cov tests/session/train/test_trajectory_analyzer_component.py tests/unit/session/memory/test_experience_lineage.py
21 passed

.venv/bin/ruff format --check openviking/session/memory/experience_lineage.py tests/unit/session/memory/test_experience_lineage.py
2 files already formatted

.venv/bin/ruff check openviking/session/memory/experience_lineage.py tests/unit/session/memory/test_experience_lineage.py
All checks passed

The focused mypy invocation traverses the repository and reports the existing main-branch type-error baseline; it reports no error in the changed lineage module.

Checklist

• My code follows the project coding style
• I have performed a human self-review of the code
• The narrow parsing boundary and existing canonical URI validation were reviewed by the AI agent
• Documentation updated (not applicable; no public API or configuration change)
• No dependency or generated-file changes

Additional Notes

The parser intentionally accepts only the exact envelope and generated list-item shape owned by skill-experience.mjs; ordinary auto-recall context and lookalike text outside that envelope are ignored. No claim of independent human review is made.

[fukalous]

Following up on the injected skill-experience lineage fix. Both current-head workflows are awaiting approval (action_required). Could a maintainer approve the runs and review the change when convenient? The parser is limited to the bundled hook's canonical envelope and retains the existing current-user Experience URI validation.

[fukalous]

Rebased cleanly onto current main (4c9a755) in 4a31b4b; no implementation conflict or follow-up code change was needed.

Fresh validation:

• trajectory analyzer + Experience lineage tests: 21 passed
• Ruff check passed
• Ruff format check passed
• git diff --check passed

[fukalous]

@heaoxiang-ai this adjusts Experience trajectory lineage for injected skill memories in the area you originally implemented. The current head is mergeable and all applicable plugin and API/CLI checks are green. Could you review the lineage behavior when convenient?