| name | canon-status |
| description | Use when a Canon run already exists and you need its current state, pending approvals, and next steps. |
Canon Status
Support State
available-now
default visibility: prominent
Purpose
Show the current Canon-backed state of an existing run.
When To Trigger
- The user has a run id and wants the latest run state.
- The user wants to know whether a run is blocked, completed, or waiting for approval.
When It Must Not Trigger
- No run id is available yet.
- The user is asking to start a new Canon workflow.
Required Inputs
Preflight Profile
- Verify
canon is on PATH. If missing, point to the install guide.
- Verify
.canon/ exists. If missing, point to $canon-init.
- If no run id was provided, resolve the latest run from Canon runtime state.
- Verify the selected run exists before invoking Canon.
- Do not show preflight checks to the user. Report only the specific missing input.
Canon Command Contract
- Canon command:
canon status --run <RUN_ID>
- Return Canon-backed state only.
Continuation Semantics
- When status surfaces
suggested_continuation, treat it as advisory only.
- Candidate detection is advisory. Continuation requires explicit intent.
refinement_state may appear for any run with persisted continuity data. Targeted planning modes can include a working-brief path, while non-targeted modes may surface continuity without a targeted working brief.
- Recommend
$canon-resume only when the user explicitly wants to continue the same run or Canon reports the run still needs continuation.
Expected Output Shape
- concise run summary
- real run id
- persisted run owner when Canon already recorded one
- real run state
- direct statement of what is wrong when the run is blocked or gated
- concrete readable artifact paths when Canon emitted them
- optional action chips mirroring the same Canon-backed next step:
Inspect evidence, Approve generation..., or Resume run only when each is valid for the active run
- ordered possible actions, not a flat menu of skills
- one recommended next step that preserves the run context
$canon-approve and $canon-resume only when Canon reports a gated state, pending approvals, or an explicit approval target
Failure Handling Guidance
- If
.canon/ is missing, point to $canon-init.
- The preflight asks only for the missing or invalid
RUN_ID and must not invent any other missing state.
- If the run id is missing or unknown, ask only for the exact run id and show the exact retry form
canon status --run <RUN_ID>.
- If Canon fails after preflight succeeds, report it as a Canon-execution outcome rather than a preflight failure.
Next-Step Guidance
- Use
$canon-inspect-artifacts as the recommended next step when Canon already emitted a reviewable packet that explains the block.
- Use
$canon-inspect-evidence when the primary need is the invocation rationale, policy decision, or evidence lineage.
- Use
$canon-inspect-invocations only when request-level detail is the next useful step, not as a default detour.
- If the run is completed and not gated, recommend the single most useful inspection surface and do not suggest
$canon-approve or $canon-resume.
- If the run is gated or awaiting approval, recommend inspection first unless the user has already reviewed the evidence; then point to
$canon-approve, followed by $canon-resume only when Canon still requires continuation.
- If no concrete artifact paths are available from Canon-backed output, do not claim that artifacts were generated and do not recommend
$canon-inspect-artifacts as the primary next step.
- If a host renders chips, keep them aligned to the same ordered next steps and never use
Proceed with generation as the approval label.
Related Skills
$canon-inspect-invocations
$canon-inspect-evidence
$canon-approve
$canon-resume