Prompt 49: Spark autostart troubleshooting should include read-only wrapper and startup-path checks first

[Aeyod7]

Problem

For Spark Compete prompt 49, Spark correctly respected the user's instruction to start with read-only checks and not change startup settings yet.

The gap is that the first autostart diagnostic response did not include the known high-value read-only checks for stale wrappers, old binary paths, moved files, or startup entries pointing at removed paths.

That matters because autostart failures are often path/wrapper issues, not just “service is down.”

Evidence

User prompt tested:

Spark autostart is not working. Troubleshoot it, but start with read-only status checks first. Do not change startup settings yet.

Observed Spark response:

Let me check what's going on with your Spark startup.

**Quick checks to run:**

1. `spark status` -- general health
2. `spark live status` -- background service state
3. `spark verify --onboarding` -- config sanity

Run those and paste the output. I'll read the signals and tell you what's blocked before touching anything.

What worked:

• Spark started read-only.
• Spark did not mutate startup settings prematurely.
• Spark gave a clear stop point before proposing changes.

Gap:

• No immediate check for stale wrapper path.
• No check for old Spark binary path.
• No check for startup entry pointing to removed/moved files.
• No WSL/systemd/session distinction.
• No distinction between “autostart not installed” and “service configured but not running.”

Sanitized local capture used for this report:

/tmp/spark-qa-user-prompts-3-4-49.raw.txt

No secrets, tokens, cookies, wallet material, or private logs are included.

Fix

Expand the first autostart troubleshooting answer to include read-only wrapper/path checks before any mutation.

Suggested first-pass checklist:

spark status
spark live status
spark verify --onboarding
spark autostart status
command -v spark
readlink -f "$(command -v spark)"

The support response should also explicitly separate these cases:

1. Spark process/service is not running.
2. Autostart is not configured.
3. Autostart is configured but points to an old/missing Spark path.
4. The shell wrapper points to an old binary.
5. WSL startup/session behavior differs from systemd/native Linux behavior.

Do not run mutating commands like enabling/disabling startup until the read-only state is known.

Verification

Reviewer can verify with this prompt replay:

Spark autostart is not working. Troubleshoot it, but start with read-only status checks first. Do not change startup settings yet.

Expected passing behavior:

• Response starts with read-only status checks.
• Response includes wrapper/path checks such as command -v spark and readlink -f "$(command -v spark)".
• Response includes autostart status/config inspection before mutation.
• Response does not enable, disable, reinstall, or rewrite startup settings in the first answer.
• Response distinguishes WSL/session startup from native service/systemd assumptions.