← Back to Hermes
Hermes recovery

Debugging and recovery

Find the failing lane first so recovery stays narrow instead of turning into a broad reinstall reflex.

Recovery ruleUse the smallest proof path first: path, model, sessions, platform behavior, and approval state before rebuilds.

When Hermes feels broken, the fastest way to lose time is to treat every symptom like a runtime failure. The current research pack supports a better rule: first decide which lane actually failed, then use the smallest proof or inspection step that can confirm it.

Start by identifying the failing lane

Use a fast triage sequence first

  1. confirm Hermes is actually installed and on PATH
  2. confirm the expected provider/model pairing is configured
  3. confirm you are in the expected session
  4. confirm the surface behavior matches the platform rules
  5. confirm approval, authorization, or backend settings did not silently change
  6. use validation and observability helpers when you need proof
  7. only then choose restart, reconfigure, or rebuild

Short recovery entries that solve a lot of cases

Related pages