req-decide

name	req-decide
description	Synthesize all evidence (CV screening, interview assessments, transcripts) into a final hiring recommendation. Produces a comprehensive recommendation document with full evidence trail, level/track/discipline confirmation, and a clear hire/no-hire decision. Use when all interview stages are complete.

Hiring Decision

Synthesize all available evidence for a candidate into a final hiring recommendation. This is the culmination of the hiring pipeline — every prior assessment feeds in.

This is Stage 3 of the three-stage pipeline:

1. req-screen — CV → interview or pass.
2. req-assess — transcript → updated evidence.
3. Hire / no hire (this skill) — all stages complete → recommendation.

Trigger

• The user asks for a final hiring recommendation.
• All planned interview stages for a candidate are complete.
• The user asks "should we hire {Name}?".
• The user needs to compare finalists for a position.

Prerequisites

• fit-pathway CLI installed.
• Knowledge/Candidates/{Name}/screening.md exists.
• At least one Knowledge/Candidates/{Name}/interview-*.md.
• Knowledge/Candidates/{Name}/brief.md exists.

Inputs

• All artifacts in Knowledge/Candidates/{Name}/.
• Standard data via fit-pathway.
• Knowledge/Candidates/Insights.md — cross-candidate context.
• Knowledge/Roles/*.md — the candidate's requisition (positions, hiring manager, domain lead). Check the **Status:** field to distinguish active from historical reqs.
• Knowledge/Priorities/*.md — strategic context.
• Other active candidates at the same level — relative positioning.

Outputs

• Knowledge/Candidates/{Name}/recommendation.md — final recommendation.
• Updated Knowledge/Candidates/{Name}/brief.md — status and link.
• Updated Knowledge/Candidates/Insights.md — cross-candidate observations.

<do_confirm_checklist goal="Verify the recommendation is grounded and audit-ready">

• Every skill rating traces to a specific evidence source (stage + detail).
• Evidence hierarchy respected — interview evidence outranks CV.
• Level recommendation grounded in standard progression criteria.
• Track recommendation cites interview evidence, not just CV.
• Decision rules applied strictly; match % and gap counts verified.
• Risk assessment is honest — concerns not minimised to justify a hire.
• "Do not hire" written with the same rigour as "Hire" (GDPR Article 15: candidate may request access).
• Brief updated with [Hiring Recommendation](./recommendation.md) and a final pipeline entry.
• Recommendation could be shown to the candidate without embarrassment.

</do_confirm_checklist>

Procedure

1. Gather all evidence

ls Knowledge/Candidates/{Name}/

Read in order: brief.md, screening.md, every interview-*.md, relevant transcript-*.md, and panel.md if present. Build the chronological evidence timeline (date / stage / source / key finding).

2. Build the final skill profile

bunx fit-pathway job {discipline} {level} --track={track} --skills
bunx fit-pathway progress {discipline} {level} --track={track}

For each skill in the target job, pick the highest-fidelity evidence using the evidence hierarchy. Record final rating, best evidence source, trajectory, confidence.

3. Build the final behaviour profile

bunx fit-pathway behaviour --list

Apply the same hierarchy (reference). Record final maturity, best evidence, consistency.

4. Confirm level and track

bunx fit-pathway job {discipline} {lower_level} --track={track}
bunx fit-pathway job {discipline} {target_level} --track={track}
bunx fit-pathway progress {discipline} {lower_level} --track={track}

Apply the level-confirmation criteria in references/rubric.md.

5. Read role context

If brief.md carries a Req:

ls Knowledge/Roles/ | grep "{req_number}"
cat "Knowledge/Roles/{matching file}"

Check the **Status:** field. Capture remaining positions, hiring manager, domain lead, goal alignment, other candidates on the same req, and channel (hr / vendor). Frame the recommendation in terms of strategic impact.

6. Assess against the active pipeline

cat Knowledge/Candidates/Insights.md

for dir in Knowledge/Candidates/*/; do
  if [ -f "$dir/screening.md" ]; then
    head -5 "$dir/screening.md"
  fi
done

Apply the pipeline-context guidance in references/rubric.md.

7. Pick the recommendation

Apply the decision rules and run the hire criteria check.

8. Write the recommendation

Save to Knowledge/Candidates/{Name}/recommendation.md using the template in references/template.md. Carry the advisory-only banner. Cite specific evidence in skill, behaviour, level, and track sections.

9. Update brief and insights

Update brief.md:

• Set Status to recommended or not-recommended.
• Append - [Hiring Recommendation](./recommendation.md).
• Add the final pipeline entry with date and outcome.

Update Knowledge/Candidates/Insights.md only when there's a cross-candidate observation worth keeping (strongest at level, sourcing channel pattern, level-adjustment implications).

Use targeted Edit operations — never rewrite entire files.