skillxp harnesses

To see where each supported harness discovers project-level skills, and which harness releases this build was validated against, use harnesses:

  $ skillxp harnesses
antigravity    validated 1.4.12    flags 1.4.12    project skills: .agents/skills   does NOT record injected context (evidence is inference)
claude-code    validated 2.1.309   flags 2.1.309   project skills: .claude/skills   records injected context
codex          validated 0.166.0   flags 0.166.0   project skills: .codex/skills    records injected context
copilot        validated 1.0.105   flags 1.0.105   project skills: .github/skills   records injected context

validated is the newest release each harness’s skill lore (discovery locations, listing evidence, resume behavior) was re-confirmed on; flags is the release the underlying agentsummons flag surface was validated on. Both record coverage, not a compatibility bound: newer releases usually keep working.

The injected-context column matters for reading results: on a harness that records injected context, discovery evidence is direct; on antigravity, it is behavioral inference. See Harness Lore.

skillxp doctor

To compare the harness versions you have installed against the validated ones, use doctor. It is free: nothing runs beyond each harness’s version command.

  $ skillxp doctor
antigravity  installed 1.4.12, validated 1.4.12 — clean
claude-code  installed 2.1.309, validated 2.1.309 — clean
codex        installed 0.167.0 > validated 0.166.0 — drift candidate; run `skillxp drift probe` to revalidate
copilot      installed 1.0.105, validated 1.0.105 — clean

A drift candidate is a statement about validation coverage, not a failure: observations still run, and a moved skill directory fails loudly (the skill never loads) rather than going unnoticed. The exit code makes doctor usable as a gate: 0 clean (harnesses that are not installed do not count), 1 at least one drift candidate, 2 a version probe failed.

skillxp observe

To stage a skill, activate it, and trace how phrases reached the model, use observe:

  skillxp observe -harness claude-code -install ./my-skill \
  -prompt "Activate the my-skill skill and follow its instructions." \
  -activation -trace "PHRASE-IN-BODY-1234,PHRASE-IN-REFERENCE-5678" \
  -out out/

observe writes a bundle:

  • observation.json: run metadata and the trace report
  • session.json: the normalized transcript (an agentminutes session record)
  • the archived native transcript(s) that evidence line numbers point into

With -runs N, each repetition gets a fresh fixture and session under run-NN/, and summary.json reports how many runs delivered each traced phrase to each location. This is the rate report for model-dependent behavior: one run is an anecdote; N runs are a rate.

The trace report

The trace report classifies each occurrence of a traced phrase:

  • harness-injected: the platform put it in front of the model (harness-push loading)
  • tool-result: the model fetched it with a tool (model-pull loading)
  • model-output: the model emitted it
  • human-prompt / echo: your own prompt, or a system record replaying conversation content; never loading evidence

Choosing phrases to trace

Never put a phrase you plan to trace in the prompt: a literal-minded model truthfully answers “it’s in your message”, and the phrase contaminates every echo location in the transcript. Seed the phrases in the skill body and reference files instead, and let the trace report show whether they arrived.

Sandboxing

-sandbox runs the harness against an isolated home cloned per run, and unlocks user-scope installs (-install-user), which are refused unsandboxed. See Sandboxing for the isolation model and per-harness auth setup.