01 / 08
Reading outputs is a skill that deepens with every system you govern
Loading…
Module A3 · 35 min
Reading agent outputs
A CAFCR helicopter view for operators: read every agent output across five layers — customer objectives, application context, functional spec, conceptual coherence, and realisation. The skill that decides whether an output is approvable, not just plausible.
Free with account
● In your library
Objectives
- Apply the CAFCR five-layer review to any agent output.
- Distinguish functional, reasoning, and confidence-vs-accuracy failures.
- Decide approve / hold / reverse / escalate using the same verdict every time.
- Produce a one-paragraph review note that survives a regulator's read.
02 / 08
CAFCR · Customer Objectives — what is this output FOR?
Loading…
03 / 08
CAFCR · Application — what is the downstream context?
Loading…
04 / 08
CAFCR · Functional — does it meet the specification?
Loading…
05 / 08
CAFCR · Conceptual — is the reasoning coherent?
Loading…
06 / 08
CAFCR · Realisation — confidence is not accuracy
Loading…
07 / 08
The three failure modes you will see most
Loading…
08 / 08
Output Reader simulation
Loading…
Recommended assets
Add human consultancy
Every tier is priced under € 99. Pick the depth you need; you’ll receive a scheduling link by email within one working day.
Consultancy · Tier 2 — Working session
60-min scoped working session + written follow-up (≤ 1 page). Free — request a session.
Free · request a session
Request a session →Consultancy · Tier 3 — Architecture review
90-min review + 2-page architecture note. Two reviewers on the call. Free — request a session.
Free · request a session
Request a session →