intelliaide: add skill metadata, OWNERS, config, and README

[sakshipatels98-byte]

Summary

Add IntelliAide as a new agentic skill for deep root-cause analysis of OpenShift
cluster issues using pre-mounted diagnostic data.

• IntelliAide RCA pipeline — SKILL.md orchestrates a 4-step workflow
  (extract_cluster → select_files → analyze_data → perform_rca) with
  3-pass priority analysis (High → Medium → Low). Python scripts handle
  computation only (token estimation, chunking, file I/O); the orchestrating
  agent performs all LLM reasoning.
• PVC-based data source — diagnostic data is read from /data/input,
  mounted read-only from the Proposal's spec.dataSource PVC. No manual
  oc cp or wait step; extract_cluster.py validates the pre-populated mount
  at pipeline start.
• Vendored Python dependencies — requirements.txt deps (requests,
  PyYAML, google-auth, drain3, odfpy, python-docx) are vendored under
  intelliaide/vendor/ for Python 3.12 so the restricted sandbox can run
  scripts without pip install at runtime.
• Supporting assets — must-gather routing documentation (DataSource/),
  ML log/YAML classifiers, and pipeline configuration (Config/).
• Repo hygiene — intelliaide/README.md documents the pipeline and vendor
  regeneration; intelliaide/OWNERS for Prow review assignment; YAML
  frontmatter on SKILL.md for skill discovery; simplified config.json
  (removed obsolete oc cp-related keys).

Test plan

• Build skills image (Containerfile) and confirm intelliaide/ is copied to /app/skills/
• Verify skill discovery: agent selects intelliaide for RCA/deep-analysis requests via SKILL.md frontmatter
• Create Proposal with spec.dataSource.claimName pointing to a PVC pre-populated with must-gather data
• Confirm /data/input is mounted and extract_cluster.py validates data without oc cp
• Run full pipeline end-to-end: extract → select → analyze → perform_rca produces structured diagnosis output
• Confirm vendored deps load correctly (no runtime pip install) inside the sandbox
• Verify config.json keys match what skill scripts consume (no references to removed must_gather_incoming_dir / must_gather_wait_seconds)
• Confirm progress annotations appear on the Proposal during pipeline execution

[openshift-ci]

Skipping CI for Draft Pull Request.
If you want CI signal for your change, please convert it to an actual PR.
You can still manually trigger a test run with /test all

[openshift-ci]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: sakshipatels98-byte
Once this PR has been reviewed and has the lgtm label, please assign cali0707 for approval. For more information see the Code Review Process.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[openshift-ci]

@sakshipatels98-byte: The following test failed, say /retest to rerun all failed tests or /retest-required to rerun all mandatory failed tests:

Test name	Commit	Details	Required	Rerun command
ci/prow/eval	46eba4f	link	true	/test eval

Full PR test history. Your PR dashboard.

Details

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. I understand the commands that are listed here.

[harche]

The PVC-based data source approach is being questioned on the companion operator PR — the sandbox pods are ephemeral and the agent can collect must-gather data itself during its run using /tmp. See this comment.

If that feedback lands, this script would need to collect the data instead of assuming it's pre-mounted at /data/input.

[harche]

This PR doesn't include evals. The repo has an eval framework under evals/skills/ — each existing skill has a directory with test_cases.yaml and system_prompt.md. Could you add evals/skills/intelliaide/ with test cases covering the pipeline?

It would be especially valuable to include negative cases that demonstrate why this skill is needed — e.g., what happens when the agent tries to analyze a large must-gather bundle without the IntelliAide pipeline (buffer overflow from raw file reads, shallow diagnosis that misses root cause, etc.). Those cases help justify the skill's existence and guard against regressions if someone later simplifies the flow.

[harche]

Individual teams are responsible for the ownership of the skill.