---
requires: objective, current evidence and feedback, and the caller's decision protocol
produces: Advance or Iterate with evidence, or a Blocked outcome for human resolution
action_style: procedural
---
Decide whether the Flow should advance, iterate, or ask a human to resolve stalled progress.
Use the objective, success conditions, and constraints supplied by the caller.
Do not impose a particular workflow, producer skill, artifact format, or
progress metric. Work and review steps supply domain-specific criteria and
evidence; this step owns the navigation judgment.
Read the previous direction, what changed or was learned, current feedback,
and unresolved findings. Follow supplied artifact paths and inspect their
current contents when they may have changed. Reconcile earlier evidence and
accepted changes; a newer proposal does not override an accepted requirement.
Reuse applicable results without treating a process exit as task completion.
An unavailable or human-dependent check alone never means Blocked. Use a
headless equivalent or name its later gate/CI owner and continue. Before gate,
build/focused-test sanity is enough. Actual failures and missing behavior still
need work. Leave judgment to demo/review; neither claim deferred checks passed
nor bypass an authored review.
- **Advance** when the work required at this boundary is satisfied, including
applicable acceptance conditions and findings from earlier attempts. Leave
later obligations to their declared steps.
- **Iterate** when meaningful progress was made and specific remaining work can
use the declared backward edge. Supply the next action and observable result,
within the accepted scope. Learning that rules out a hypothesis or narrows
the problem counts as progress; activity alone does not.
- **Blocked** when the previous iteration made no meaningful progress, repeated
the same failure without new evidence, or exposed an input or judgment the
agent cannot supply. Honor unresolved blockers in the supplied criteria and
evidence. Explain what needs resolving. Do not spend another iteration repeating unchanged direction.
If evidence is missing, name that gap rather than pretending it establishes no progress.
Separate the questions before deciding: which obligations are satisfied, what
changed or was learned, and what useful action or missing human input remains.
Inspect the relevant evidence; do not substitute a confidence score for observed results.
The runner owns edge targets and execution identity. There is no pass limit;
judge whether to continue from the work and evidence, not the iteration count.
Use the supplied decision protocol for this exact occurrence. Advance and
Iterate navigate; Blocked reports a stopped execution requiring an Ask. Include
the attempted direction, before/after evidence, and the question in that outcome.
Let the supplied protocol own how help is requested; do not open a second
Session alongside its request.
After the human completes the Ask, read its summary and the changed artifacts.
Reassess using the new evidence or direction. Completion of an Ask does not establish
that the work is done and does not choose a navigation decision. If the
same blocker remains unresolved, report that fact instead of cycling through
identical Asks automatically.
The same applies to other reviews: completion returns feedback, not a verdict.
Carry accepted changes, unresolved questions, and the concrete next action into
the decision summary, with references the next step can follow.
A rejected decision command is correction feedback, not a failed work pass.
Correct its reported shape or authority error within this Run when possible;
do not rerun implementation to repair output. Missing or malformed output does
not establish that the work made no progress. If the protocol still cannot be
satisfied, leave the specific failure visible rather than inventing success.
Without a bound decision protocol, return this assessment in the conversation.
Do not invent a decision command, create a Task, launch a successor, or write a
file for a later agent to interpret as execution authority.