Skip to main content

run-a-candidate-debrief

Facilitate a candidate debrief that reconciles independent evidence against predefined role criteria. Use when a hiring team must make or recommend a selection decision.

Source facts

Repository
protosphinx/sphinxstack
Last source activity
July 29, 2026 at 13:55
Detected SKILL.md language
English
Stars
3
Forks
0

Install options

The review-first prompt is selected by default. You can switch to a direct command or download a local copy.

Review the source files

Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.

File Explorer
2 files

Showing SKILL.md

SKILL.md
Source instructions · Read-only preview
name
run-a-candidate-debrief
category
get-hired
description
Facilitate a candidate debrief that reconciles independent evidence against predefined role criteria. Use when a hiring team must make or recommend a selection decision.
# run-a-candidate-debrief Preserve independent observations before discussion. Resolve evidence gaps without turning confidence, seniority, or vague fit into proof. ## Procedure 1. Confirm the role, decision authority, scorecard, candidate stage, conflicts, and required attendees. 2. Require assessors to submit ratings and evidence before seeing group sentiment. 3. Start with must-have gates, assessment integrity, accommodations, and missing evidence. 4. Review each criterion with specific observed behavior, work sample, or verified record. 5. Separate fact, inference, prediction, and preference; ask unsupported judgments to be withdrawn or evidenced. 6. Examine rating differences for distinct observations, anchor interpretation, or evaluation bias. 7. Compare the candidate with the role standard, not with demographic or stylistic similarity to the team. 8. Decide advance, hold for bounded evidence, reject, or recommend offer using the approved rule. 9. Record rationale, dissent, evidence gaps, candidate communication owner, and any process issue. ## Failure plan - Do not reveal irrelevant protected information or let it influence the decision. - Never backfill scorecards to match a group favorite. - If the process treated candidates inconsistently, pause and repair comparability before selection. - Avoid collecting extra references or work selectively to rescue only one candidate. ## Done - Independent scores, evidence, differences, and missing data are reconciled - The decision follows the predefined role criteria and authority - Rationale, dissent, communication, and process corrections are recorded
View on GitHub