Lost Search Post-mortem skill

What the Lost Search Post-mortem skill does

Best for

Agency recruiters and owners reviewing a search that was lost or dragged on

Author

Work Insiders

Needs

Nothing beyond your AI assistant

License

MIT

Runs in

Claude, Gemini Enterprise, Gemini CLI, Gemini app (as a Gem), ChatGPT, Codex, Grok

Nothing hidden: this is the whole file. Copy it into a Gem or GPT, or download the ZIP for Claude and Gemini Enterprise. Want it with your agency's details already filled in, or formatted for a specific app? Use the copy tool.

---
name: lost-search-post-mortem
description: Reviews a lost, cancelled or slow search to find why it went wrong and what to change on the next job order, from intake to offer. Use when a recruiter says "post-mortem", "why did we lose this", "the search died", "another agency beat us", "the offer fell through", or at the end of any search that took too long.
license: MIT
metadata:
  author: Work Insiders
  version: "1.0"
  pack: agency-desk-pack
---

# Lost search post-mortem

Every lost search teaches something, but only if someone writes it down while it is fresh. Your job: find the one or two moments that decided the outcome and turn them into changes for the next job order. No blame. The goal is the next placement.

## Input

Whatever exists: the job order, dates of key events, the submittal list with outcomes, client and candidate feedback, emails. Messy notes are fine.

## Step 1: the timeline

Lay the search out as dated steps: job order taken, first submittal, each interview, feedback, offer, end. Mark any gap longer than five working days. Where a date is missing, write `date unknown`.

## Step 2: the numbers

Where the notes allow, show:

- Days from job order to first submittal.
- Submittals, first interviews, final interviews, offers.
- Interviews per submittal.
- Average days the client took to give feedback.

Do not calculate anything the notes do not support. Say what is missing.

## Step 3: find where it broke

Mark each stage OK, Weak or Broke, with the evidence:

| Stage | Questions to answer |
|---|---|
| Intake | Was the job order scored? Were must-haves capped at three? Did we talk to the hiring manager? Was the range realistic? |
| Calibration | Did we calibrate before the full search? Did the brief change afterwards? |
| Sourcing | Did we only work the obvious pools? How long until the first strong profile? |
| Submittals | Did rejections share a reason? Did we submit too few, too many or too late? |
| Client process | Slow feedback, extra rounds, new decision makers, changing requirements? |
| Candidate side | Drop-outs, counter-offers, competing offers, surprises on pay? |
| Competition | Other agencies, an internal hire, the role pulled? |

## Step 4: the verdict

- **Main cause:** one sentence. Pick one; a list of ten causes teaches nothing.
- **Contributing cause:** one, if there is a real second one.
- **Within our control?** Yes / Partly / No. Be honest: some searches die because the company froze hiring.
- **Early warning:** the first sign this was going wrong, and when it showed.

## Step 5: changes

Write two to four specific changes, each tied to a stage and a skill or habit. Examples of the right shape:

- "Intake: renegotiate or walk away when the job order scores under 5 (Job Order Intake skill)."
- "Calibration: send three calibration profiles before the first full submittal with every new client (Role Calibrator skill)."
- "Offer: pre-close pay before final interviews when the gap is over 10% (Offer Pre-close skill)."

Avoid vague lessons like "communicate better".

## Step 6: a note to the client (optional)

If the relationship is worth keeping, draft a short, gracious note: thanks, one honest line on what you learned, and an offer to help with the next opening. No excuses and no blaming the client.

## Output

```
POST-MORTEM: [Role] at [Client] | Result: [lost to / cancelled / filled late]
Timeline: [key dates]
Numbers: [ ]
Main cause: [ ] | Contributing: [ ] | In our control: [ ]
Early warning: [ ]
Changes for next time: 1. 2. 3.
Client note: [optional]
```

## Rules

- Write about the process, not people's character.
- Leave out candidates' personal details. Use initials.
- Do not guess at the client's internal reasons. Mark them `unknown` unless someone told you.

# Agency profile

Fill this in once. Every skill in the pack reads it, so outreach, job orders and submittals come out in your agency's voice with your real terms. Delete any line you do not want the AI to use. Leave a line blank rather than guessing.

## Who we are
- Agency name:
- Recruiter name and title (for sign offs):
- Phone:
- Mailing address (needed for US commercial email):
- Years in business:
- Specialties (roles we fill):
- Markets (cities, regions or "US remote"):
- Industries we know best:

## Terms
- Direct hire fee basis (e.g. % of first-year base):
- Guarantee period:
- Contract markup or bill rate approach:
- Exclusive or retained options:

## Proof points we can share
- Placements, time to fill, or retention figures we are happy to quote:
- Client names we may mention (only with permission):

## How we work
- ATS:
- Do we text candidates? (yes/no):
- Name the client in first candidate messages? (yes/no):
- Blind profiles: always / new clients only / never:
- Include pay expectation in submittals? (yes/no):
- Job post apply instructions:
- EEO statement:

## Words we never use
-

How to install Lost Search Post-mortem in Claude, ChatGPT, Gemini or Grok

Steps follow OpenAI's skills guide and Google's Gemini Enterprise and Gemini CLI guides. Menu names change, so check there if a label looks different. Claude Code reads skills from ~/.claude/skills/.