agentleFS
timscheueraitimscheuerai/gtm-architecture-template/.agents/skills/oxygen-diagnostics/SKILL.md

oxygen-diagnostics

Skill2 starsChanged 2 months agoMITView on GitHub

Trace OXYGEN cells and runs, explain failures or credit spend, and plan reconciled retries from native provenance.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read the runtime contract.

    line 8
  2. Distinctive:

    Use run and item provenance to connect the symptom to its owner.

    line 19
  3. Distinctive:

    Before a retry, identify exact items, possible duplicate effects and incremental cost.

    line 21

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Explain one operational failure

Read the runtime contract. Identify the visible Table/row/cell, Workflow, Sequence or message and the user's symptom. Discover oxygen capabilities search "diagnose failed table run" --json using the actual symptom, then hydrate the narrow read command.

QuestionEvidence to collect
Wrong or missing cellInput values, dependencies, column definition/version, cell/run envelope, sources
Failed or stuck workStatus, attempts, lease/retry time, structured error, approval/integration state
Unexpected creditsRow count, provider/model operation, attempts, caps, estimated versus charged receipts
Ambiguous external effectNative action/message ID, provider ID, delivery state and reconciliation result

Distinguish pending, leased, deferred, approval-required, retryable and failed work. Missing output may mean the step never ran; an empty view may be scoped or stale. Read envelopes and timestamps instead of assuming a raw string is the final result.

Use run and item provenance to connect the symptom to its owner. Avoid dumping every workspace log or joining unrelated runs solely by time. A cost estimate, reservation and settled charge are not interchangeable.

Before a retry, identify exact items, possible duplicate effects and incremental cost. Reconcile an ambiguous send or write first. Preview the bounded repair and execute only within the authorized scope; do not switch providers or build hidden local recovery state.

Return the most likely cause, supporting IDs/links, remaining uncertainty and verified repair state. If a product defect remains, prepare a reproducible report locally; sending feedback or transcripts requires the user's request.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.