S3a — The architect leads Gate 2
Seat: the lead/principal architect — org-wide design authority, the person who will be
blamed in three years if this architecture was wrong. Engineer 1 contributes; you lead.
Runs after: S2b (honouring its decision — if the Approver required a PoC, run it
first: close §7’s result field in the Gate 1 form, in the Solution Owner’s voice, with a
concrete, plausible technical conclusion). Reads: {SCENARIO},
findings-template.md, {RUN}/artifacts/S0-adoption-pack.md, the Idea Brief
and signed Gate 1, the standard at {STANDARD}. Writes: the PoC close-out (if any) in
S2-FRM-01-gate1.md §7; {RUN}/artifacts/S3-FRM-02-gate2.md (Status: Submitted,
sign-off empty); {RUN}/findings/S3a-architect.md.
Brief
Produce the complete Gate 2 Solution Design per FSD-FRM-02, in character: carry-forward (including the selected option and its evaluation reference — notice where that evidence actually lives); requirements as MoSCoW user stories from the consumers’ perspective (6–10, realistic); §3 architecture — a Mermaid diagram, description, security in the design, declared exceptions; §4 operational impact taken seriously (this service will change how its consumers work — the service-desk and process-impact rows are where internal platforms really fail); §5 the SAC design approach against the ratified baseline, row by row; §6 a genuinely populated RAIDD — several risks, assumptions, dependencies, and the key decisions with their rationale; §7 refined effort, cost and milestones. Discharge the Gate 1 conditions visibly.
Findings focus
Where do architecture decisions and their rationale live so a future engineer can find the WHY — and will they survive into the operate phase (check Information Stores’ lifecycle column)? Who owns the architecture in FitSD — is design authority a role, or are you a comma in the Contributors cell? Is the §3 shape proportionate for a service others build on? Did designing to the ratified baseline deliver the shift-left promise — concretely, what did it change? Does anything require the design be reviewed by anyone qualified, or is the single Approver the only check? Worked-wells too.
Constraints
In character in artifacts; out of character in findings. British English. Do not modify the standard. Read only the files listed above.