Record an attempt with observations and a next practice task

After a practice session, record what happened while it is still fresh. For example, write “the second delivery changed the total from 2 to 4” instead of “weak on idempotency.” Link the code or drawing, the input that exposed the problem, and the next attempt. Copy the form below for each session.

Use one record per occasion; keep the assessor's raw observations, not only a total. The rubric defines the curriculum gate and its limits.

Date / candidate / assessor / intended scope (senior, team lead, organization lead):
Session / variant / candidate confirms unseen material:
Mode and allowed tools (independent, or explicitly AI-permitted):
Start / follow-up release / finish timestamps:
Candidate-visible files and held-back materials released:
Contract stated and first concrete worked example:
Runnable implementation or design artifact and exact commands:
Failed invariant / smallest reproduction / repair and regression:
Clarification 0–3 + observation:
Correctness 0–3 + observation:
Architecture 0–3 + observation:
Implementation 0–3 + observation:
Tradeoffs 0–3 + observation:
Operations/security 0–3 + observation:
Communication 0–3 + observation:
Ownership/influence 0–3 + observation:
Unfinished scope; prompts supplied by assessor:
Debrief prerequisite / next unfamiliar variant / revisit date:
Other independently assessed occasion linked:
Curriculum gate met or next specific repair; no hiring probability inferred:

A role dimension outside the prompt is not assessed, with a reason; do not silently invent a score. Use the worked scored examples to calibrate what “independent” and “with prompting” mean before conducting the session.