A Claude Skill that takes a job profile plus an ICP rubric, builds an AI sourcing query against Juicebox, hireEZ, or LinkedIn Recruiter, retrieves up to 200 candidates, scores each against the rubric with cited evidence, and drafts personalized outreach for the top-N — then stops at a human-review gate. The recruiter reads the shortlist, edits the messages, and sends. Replaces the 3-hour Boolean-plus-scoring-plus- outreach loop with a 30-minute review loop.
When to use
You source for a role that runs more than once a quarter and the ICP rubric is stable enough to write down.
You have an ICP rubric with behavioral anchors per dimension (not just vague labels). The rubric template in the bundle’s references/1-icp-rubric-template.md shows the shape; if you cannot fill it in, you do not yet have a rubric this skill can score against.
You have API access to Juicebox PeopleGPT, hireEZ, or LinkedIn Recruiter. The skill refuses to fall back to scraping public LinkedIn URLs.
A human recruiter or sourcer reviews every shortlist before any outreach is sent. The skill writes drafts to disk and stops.
When NOT to use
Auto-rejection in the loop. The skill ranks; it does not reject. The “skipped” candidates are surfaced with reasons for the recruiter to overrule. Wiring a reject action to a score threshold turns this into automated decision-making and triggers EU AI Act Annex III high-risk obligations plus NYC LL 144 bias-audit obligations within one year before use. If you need that, get a bias audit, not this skill.
Scoring on protected-class proxies. School-tier as a standalone dimension, name origin, photo presence, employment-gap penalties, age inferred from graduation year, “culture fit” without behavioral anchors. The skill’s fairness checklist refuses to run if any of these appear in the rubric. Do not edit the checklist to make a biased rubric pass.
Pay-band recommendations. NYC LL 32-A, Colorado, California, and Washington require posted ranges and bias-audit obligations on automated pay decisions. Use a comp benchmarking tool, not a sourcing skill.
One-off C-suite searches. A retained search for a specific named individual or a tightly-defined exec is faster done by a human with network. The skill is built for repeatable IC and manager-level sourcing, where rubric calibration earns back its setup cost.
Reference checks or backchannel research. Different consent posture. Different workflow.
Setup
Drop the bundle. Place apps/web/public/artifacts/candidate-sourcing-claude-skill/SKILL.md into your Claude Code skills directory (or claude.ai custom Skills).
Fill in the rubric. Copy references/1-icp-rubric-template.md to a per-role file under your own repo. Replace every {placeholder}. The skill captures the rubric’s SHA-256 in its audit log per run, so subsequent edits are visible in retro.
Configure the source channel. Add your Juicebox or hireEZ API key to the skill’s config. For LinkedIn, configure Recruiter API credentials — the skill refuses to scrape public profile URLs.
Write the do-not-poach and exclude lists. A CSV of customer domains (do-not-poach) and a CSV of exclude_list URLs (recently rejected, in silent period, opted out). The deterministic pre-filter in step 3 of the skill applies these before the LLM sees any candidate.
Dry-run on a closed role. Run on a role you sourced manually last quarter. Compare the skill’s top-25 to your manual top-25. Tune the rubric anchors if the skill calibrates differently — the anchors, not the search query, are usually wrong.
What the skill actually does
Six steps, in order. The order matters: deterministic filters and fairness pre-flight come before the LLM ranking, because letting an LLM loose on a contaminated pool produces fast, confident, unusable output.
Validate the rubric against references/2-fairness-checklist.md. Halt if the rubric contains protected-class proxies. The choice to fail before retrieval rather than after is deliberate — a biased rubric loaded into a sourcing tool’s API leaves a log entry that already counts as automated processing under GDPR Art. 22.
Build the search query in the channel’s native format. Cap synonyms at 5 per dimension; cap retrieved pool at 200. Larger pools degrade ranking because the model context fills with low-relevance candidates.
Deterministic pre-filter. Drop exclude_list matches, do-not-poach companies, location mismatches, and profiles >18 months stale. These are auditable filters; the LLM does not re-litigate them.
Rubric-based ranking. Score 1-5 on skill, level, company-pattern, response-likelihood. Every score above 1 cites a verbatim profile string. No citation → score 1. The citation requirement is what keeps the model grounded in profile text rather than inferring from name, photo, or school.
Human-review gate. Write shortlist.md and per-candidate outreach/<id>.md files. Stop. The skill defines no send action.
Audit log. Append one JSONL line per run with run_id, rubric_sha256, pool sizes, channel, model. No PII. This is what makes the run defensible under NYC LL 144 or EU AI Act questioning.
The shortlist format and per-candidate evidence layout live in references/3-shortlist-format.md in the bundle. The format is fixed because downstream consumers — recruiter, hiring manager, audit reviewer — need predictable columns.
Cost reality
Per shortlist of 25 from a 200-candidate pool, on Claude Sonnet 4.5:
Retrieval cost — depends on channel. Juicebox PeopleGPT counts against your monthly query quota (200-search starter plans cap fast if you run multiple roles per week). hireEZ unlocks-per-month is the binding constraint there. LinkedIn Recruiter API has its own per-seat InMail and search quotas. None of this changes with the skill in the loop; you spend the same channel quota you would have spent doing manual Boolean.
LLM tokens — typically 80-120k input tokens (rubric + 200 candidate profile excerpts + skill instructions) and 8-15k output tokens (shortlist + 25 outreach drafts). On Sonnet 4.5 that is roughly $0.50-0.80 per shortlist. The full month for a sourcer running ~80 shortlists lands at $40-65 in model cost.
Recruiter time — the win is here, not in model cost. Manual Boolean + scoring + outreach for 25 candidates is 2-3 hours. Reviewing the skill’s shortlist and editing the drafts is 25-40 minutes, which is what makes the workflow worth running.
Setup time — 45 minutes for the rubric and exclude lists if the rubric already exists in some form; longer if the rubric is net-new (in which case structured interviewing is the prerequisite, not this skill).
Success metric
Track three numbers per role per month, in the ATS:
Reply rate to outreach — should match or exceed the recruiter’s baseline manual rate. If it drops, the outreach drafts are generic — usually the rubric is too coarse, not the model.
Shortlist-to-screen pass rate — the share of shortlisted candidates the hiring manager agrees are worth a screen. Should be ≥70% on a stable role. Below that, the ICP rubric is miscalibrated; re-run on a closed role and tune.
Time-from-role-open to first qualified screen — the throughput metric the skill is meant to move. The 3-hour-to-30-minute reduction shows up here, not in model spend.
vs alternatives
vs Gem AI Sourcing — Gem owns the recruiter workflow end-to-end (sourcing UI, sequences, analytics, ATS integration via Ashby and others). Pick Gem if you want a managed product and your team will live in its UI. Pick this skill if you want the rubric, pre-filter logic, and audit log in your own repo, version-controlled, with the model swappable.
vs hireEZ’s built-in AI ranking — hireEZ’s AI Match is good retrieval; the gap is at the rubric layer. With this skill you keep hireEZ as the retrieval channel and bring your own rubric + evidence-cited scoring on top. If hireEZ’s defaults match your ICP, you do not need this skill.
vs manual Boolean + spreadsheet scoring — manual is right for one-off or executive searches, where the rubric is in the recruiter’s head and writing it down is overhead that does not pay back. The skill earns its setup cost on roles that repeat.
vs DIY Python script against the LinkedIn / Juicebox APIs — same ranking quality if you build the prompt carefully, but you also build the fairness checklist, the audit log, and the human-review gate yourself. The bundle ships them.
Watch-outs
Bias amplification — guarded by the fairness checklist in references/2-fairness-checklist.md, which halts the run if the rubric contains protected-class proxies. The audit log captures rubric_sha256 per run so the rubric used on a given date is reproducible under EU AI Act or NYC LL 144 review.
Stale LinkedIn / Juicebox data — guarded by the deterministic filter in step 3 (drop profiles >18 months stale) and by the response-likelihood dimension in scoring (which weights freshness). Cold-storage candidates do not crowd out actively looking ones.
LinkedIn ToS exposure — guarded by refusing to scrape public profile URLs. The skill uses the Recruiter API, Juicebox, or hireEZ, which carry their own data licensing. If linkedin_recruiter is selected and the API is not configured, the skill aborts with a setup-error rather than falling back.
Auto-send drift — guarded by the human-review gate (step 5) and by the absence of any send action in the skill. Drafts are written to outreach/<id>.md files for the recruiter to paste into the ATS / sourcing-tool outbox. AI-drafted-and-sent without review produces volume without quality and damages candidate experience.
Comp transparency — outreach drafts never quote a number; they reference the band as “competitive range disclosed on screen” so the recruiter remains the source of pay-band statements (NYC LL 32-A, Colorado, California, Washington pay-transparency requirements).
Stack
The skill bundle lives at apps/web/public/artifacts/candidate-sourcing-claude-skill/ and contains:
SKILL.md — the skill definition
references/1-icp-rubric-template.md — fill in per role
references/2-fairness-checklist.md — pre-flight checks (do not edit to make biased rubrics pass)
references/3-shortlist-format.md — the literal output format
Tools the workflow assumes you already use: Claude (the model), Juicebox or hireEZ (the retrieval channel), Ashby (the ATS for write-back once the recruiter has approved a candidate). Gem is the build-vs-buy alternative if you do not want to own the rubric and audit log yourself.
---
name: candidate-sourcing
description: Translate a job profile and ICP rubric into a sourcing query, retrieve candidates from Juicebox / hireEZ / LinkedIn Recruiter, score them against the rubric, and draft personalized outreach for the human reviewer to approve. Always stops at a human-review gate before any outreach is sent.
---
# Candidate sourcing
## When to invoke
Use this skill when a recruiter or sourcer hands you a role plus an ICP rubric and wants a ranked, evidenced shortlist with draft outreach. Take a job profile (title, level, must-have skills, location, comp band) and a fairness-aware rubric as input, and produce a Markdown shortlist plus a folder of draft messages.
Do NOT invoke this skill for:
- **Automated rejection.** This skill ranks; it never rejects. The "below threshold" tail is surfaced for the recruiter, who decides. Auto-reject in the loop triggers EU AI Act high-risk obligations and most US state hiring-AI laws.
- **Scoring against protected-class proxies.** Do not ask the skill to score on "culture fit", name origin, school prestige as a standalone signal, photo, age inferred from graduation year, gender inferred from pronoun usage, or pregnancy/parental status inferred from gaps. If the rubric contains any of these, refuse and surface the rubric line for the user to fix.
- **Pay-band recommendations.** NYC LL 144, Colorado, California, and Washington require posted ranges and bias audits for automated decisions on pay. Use a comp benchmarking tool, not this skill.
- **Reference checks or backchannel research on named individuals.** That is a different workflow with its own consent posture.
## Inputs
- Required: `job_profile` — path to a Markdown file with title, level, must-have skills, nice-to-have skills, location / remote policy, comp band, and the EEOC job category.
- Required: `icp_rubric` — path to the rubric file under `references/`. Without this the skill refuses to run; an unfaitened rubric is the most common cause of biased shortlists.
- Required: `source_channel` — one of `juicebox`, `hireez`, `linkedin_recruiter`. Do not mix channels in a single run; per-channel ToS and rate limits differ.
- Optional: `n` — shortlist size, default 25, hard max 100. Above 100 the skill warns that human review will not be meaningful.
- Optional: `exclude_list` — path to a CSV of `do_not_contact` emails or LinkedIn URLs (do-not-poach customers, prior rejects within 6 months, silent-period candidates).
## Reference files
Always read these from `references/` before doing any retrieval. Without them the shortlist is uncalibrated and the fairness guards are absent.
- `references/1-icp-rubric-template.md` — the rubric the skill scores against. Replace the template content with your role-specific rubric before running.
- `references/2-fairness-checklist.md` — pre-flight checks the skill runs on the rubric and on the retrieved pool. Fail-loud if any check fails.
- `references/3-shortlist-format.md` — the literal output format, including the evidence and source-URL columns the recruiter needs to defend the shortlist downstream.
## Method
Run these six steps in order. Steps 1-3 are deterministic filters and fairness pre-flight; only step 4 uses the LLM for ranking. The order is deliberate — running the LLM over an unfiltered, ToS-violating, or rubric-contaminated pool produces output that is fast, confident, and unusable.
### 1. Validate the rubric
Open `icp_rubric` and run every check in `references/2-fairness-checklist.md`. If any line in the rubric matches a protected-class proxy pattern (school-tier scoring, name-based filtering, employment-gap penalties, photo presence, "culture fit" without behavioral anchors), stop and return the offending lines to the user. Do not proceed with retrieval.
The choice to fail before retrieval rather than after is intentional: a biased rubric loaded into a sourcing tool's API leaves a log entry that counts as automated processing under GDPR Art. 22 and the EU AI Act, regardless of whether the skill ever shows the user the result.
### 2. Build the search query
Translate the job-profile must-haves into the channel's native query format:
- `juicebox` → natural-language PeopleGPT prompt, with location and level filters set as structured parameters not free text.
- `hireez` → Boolean string with explicit AND/OR/NOT grouping. Cap synonyms at 5 per dimension; longer Boolean degrades hireEZ's relevance ranking.
- `linkedin_recruiter` → use the Recruiter API with structured filters only. **Do not scrape `linkedin.com/in/` URLs** — that violates LinkedIn ToS and the *hiQ v. LinkedIn* settlement does not change ToS exposure for production sourcing.
Cap the retrieved pool at 200. Larger pools degrade rubric scoring because the LLM context fills with low-relevance candidates and the ranking flattens.
### 3. Deterministic pre-filter
Before the LLM sees any candidate, apply hard filters:
- Drop anyone in `exclude_list`.
- Drop anyone whose current company is on the do-not-poach list.
- Drop anyone whose profile was last updated more than 18 months ago (LinkedIn / Juicebox staleness signal).
- Keep only candidates whose stated location matches the role's location policy (with a configurable radius for hybrid roles).
These filters are deterministic so they can be audited. The LLM does not re-litigate them in step 4.
### 4. Rubric-based ranking
For each remaining candidate, score 1-5 on each rubric dimension (skill-match, level-fit, company-pattern-fit, response-likelihood). For every score above 1, cite the specific evidence string from the candidate's profile. No evidence string → score 1 by default.
Why a citation requirement: it forces the model to ground each score in profile text rather than infer from a name, photo, or school. Scores without evidence are the mechanism by which bias enters AI-augmented sourcing pipelines.
### 5. Human-review gate
Stop. Write the shortlist to `shortlist.md` per the format in `references/3-shortlist-format.md`. Write the draft outreach to `outreach/<candidate-id>.md`, one file per candidate. Do not call any "send" endpoint. Do not mark candidates as contacted in the ATS. Surface the path to both directories and exit.
The recruiter's job from here: read the shortlist, edit the messages, and send through the ATS or sourcing tool's outbox. The skill does not re-enter the loop until the next role.
### 6. Audit log
Append a single line to `audit/<YYYY-MM>.jsonl` containing: `run_id`, `role`, `rubric_sha256`, `pool_size_pre_filter`, `pool_size_post_filter`, `shortlist_size`, `channel`, `model_id`, `timestamp`. Do not log candidate PII to this file. The audit log exists so that under NYC LL 144 or EU AI Act questioning, the recruiter can demonstrate which rubric was used on which date.
## Output format
```markdown
# Sourcing shortlist — {Role title}
Generated: {ISO timestamp} · Channel: {channel} · Pool: {pre} → {post} · Rubric SHA: {short}
| # | Name | Current role | Current company | Skill | Level | Pattern | Response | Aggregate | Source |
|---|---|---|---|---|---|---|---|---|---|
| 1 | Jamie L. | Senior Backend Engineer | Acme Fintech | 5 | 5 | 4 | 4 | 18 | {URL} |
| 2 | ... | ... | ... | ... | ... | ... | ... | ... | ... |
## Evidence — top 5
### 1. Jamie L. (aggregate 18)
- **Skill (5)**: "5y Go, 2y Rust, led migration from monolith to event-driven services" — profile, role 2.
- **Level (5)**: "Senior IC, scope across two teams, mentors three engineers" — profile, current role.
- **Pattern (4)**: "Stripe → Plaid → Acme Fintech" — three fintech roles in sequence.
- **Response likelihood (4)**: profile updated 11 days ago, "open to opportunities" tag set.
### 2. ...
## Skipped — surfaced for review (not auto-rejected)
| Name | Reason |
|---|---|
| ... | "current company on do-not-poach list (Acme Customer)" |
| ... | "profile last updated 2023-11, staleness > 18mo" |
## Draft outreach
Drafts written to `outreach/`. Recruiter reviews and sends; this skill
does not contact candidates.
- `outreach/jamie-l.md`
- `outreach/...`
```
## Watch-outs
- **Bias amplification (NYC LL 144, EU AI Act, EEOC).** *Guard:* the fairness checklist in `references/2-fairness-checklist.md` runs in step 1 and refuses retrieval if rubric contains protected-class proxies. Audit log in step 6 stores `rubric_sha256` so the rubric used on a given run is reproducible.
- **LinkedIn ToS exposure.** *Guard:* skill uses the Recruiter API (or Juicebox / hireEZ which carry their own data licensing), never scrapes public LinkedIn pages. If the channel is `linkedin_recruiter` and the Recruiter API is not configured, the skill aborts with a setup-error rather than falling back to scraping.
- **Stale profile data.** *Guard:* deterministic filter in step 3 drops candidates with `profile_updated > 18mo`. Response-likelihood scoring in step 4 weights profile freshness explicitly so cold-storage candidates do not crowd out actively looking ones.
- **Auto-send drift.** *Guard:* skill stops at the human-review gate in step 5 and writes to `outreach/` files. There is no `send` action defined anywhere in this skill. To send, the recruiter pastes into the ATS / sourcing tool outbox.
- **Rubric drift mid-search.** *Guard:* `rubric_sha256` is captured per run; if the rubric changes between two runs for the same role, the audit log shows both hashes, making it visible in retro.
- **Compensation discussion in draft outreach.** *Guard:* outreach templates in this skill never quote a number; they reference the comp band as "competitive range disclosed on screen" so the recruiter remains the source of pay-band statements (NYC LL 32-A, CO, CA, WA pay-transparency posting).
# ICP rubric — TEMPLATE (per role)
> Replace this template's contents with the rubric for the specific role.
> The candidate-sourcing skill scores against the four dimensions below.
> Each dimension MUST have behavioral anchors — vague labels ("senior")
> without anchors produce noisy and biased scoring.
## Role identity
- **Title**: {e.g. Senior Backend Engineer, Platform}
- **Level**: {IC4 / IC5 / EM1 — your internal scale}
- **Location policy**: {remote-US / hybrid-NYC-2dpw / onsite-Berlin}
- **EEOC job category**: {2 — Professionals (most engineers); see EEO-1}
- **Comp band (recruiter-internal, never sent to skill output)**: {range}
## Dimension 1 — Skill match (1-5)
The candidate's profile shows direct experience with the must-have technologies and the specific problem-shape of the role.
| Score | Anchor |
|---|---|
| 5 | Held a role doing exactly this work for ≥2 years; cites artifacts (talks, OSS, posts). |
| 4 | Held a role doing exactly this work for ≥1 year; no artifacts. |
| 3 | Adjacent work (e.g. Java backend role for a Go role); transferable. |
| 2 | Tangential work; would require ramp. |
| 1 | No evidence in profile. |
## Dimension 2 — Level fit (1-5)
The candidate's stated scope and tenure pattern match the level the role is hiring at. Do NOT use school prestige, employer prestige, or title inflation as a level signal — anchor on scope description.
| Score | Anchor |
|---|---|
| 5 | Profile shows scope at or above target level (multi-team, mentoring, technical strategy). |
| 4 | Scope at target level for ≥1 year. |
| 3 | One level below target; growth trajectory plausible. |
| 2 | Two levels below; reach. |
| 1 | More than two levels off, in either direction. |
## Dimension 3 — Company-pattern fit (1-5)
The shape of the candidate's prior employers matches the shape of yours (stage, scale, regulated/unregulated, B2B/B2C). Anchor on *characteristics*, not brand names — brand-name scoring is the most common bias vector in AI-augmented sourcing.
| Score | Anchor |
|---|---|
| 5 | ≥2 prior employers match {stage/scale/domain pattern}. |
| 4 | 1 prior employer matches; others adjacent. |
| 3 | All adjacent (different domain, similar stage). |
| 2 | Mostly mismatched; one transferable role. |
| 1 | No pattern match. |
## Dimension 4 — Response likelihood (1-5)
How likely the candidate is to respond to outreach right now.
| Score | Anchor |
|---|---|
| 5 | Profile updated <30 days; "open to opportunities" set; recently posted about job search. |
| 4 | Profile updated <90 days. |
| 3 | Profile updated <180 days. |
| 2 | Profile updated <12 months. |
| 1 | Stale profile (>12 months) — *also flagged in pre-filter for drop at >18mo*. |
## Disqualifiers (deterministic, applied in step 3 of the skill)
These cause the candidate to be surfaced in the "skipped" table, not auto-rejected. The recruiter decides.
- Current company is on do-not-poach list (`{path-to-list}`).
- Email or LinkedIn URL appears in `exclude_list`.
- Stated location does not match role's location policy + radius.
- Profile last updated >18 months ago.
## Bias guards (refusal triggers — skill aborts in step 1 if present)
If any of the following appear in this rubric, the skill refuses to run:
- School-tier scoring as a standalone dimension.
- Name-based filtering or scoring.
- Photo-based scoring.
- Employment-gap penalties without a job-related justification.
- Age inferred from graduation year used in any dimension.
- Gender, ethnicity, religion, sexual orientation, parental status, or disability status as a scored or filtered dimension.
- "Culture fit" without behavioral anchors.
## Last edited
{YYYY-MM-DD} — bump on every material change. The skill captures the SHA-256 of this file in its audit log per run.
# Fairness pre-flight checklist
> The candidate-sourcing skill runs every check below in step 1 (rubric
> validation) and step 3 (post-filter pool review). Any failed check
> halts the run with a message naming the failure. Do not edit this file
> to make checks pass — fix the rubric or the search instead.
## A. Rubric checks (run before retrieval)
A1. **No protected-class proxies.** Scan the rubric for any of the following terms or patterns. Any hit halts the run:
- `school`, `university`, `Ivy`, `tier-1`, `top-N` (when used as a scoring dimension, not as one signal among many)
- `name origin`, `surname`, `first name`
- `photo`, `headshot`, `appearance`
- `age`, `years since graduation`, `birth year`
- `gender`, `pronoun`, `she/her`, `he/him` (as filter terms)
- `ethnicity`, `race`, `nationality` (except where required for immigration-status filtering with documented legal basis)
- `pregnant`, `parental`, `maternity`, `paternity`
- `disability`, `accommodation`
- `religion`, `political`, `marital`
- `culture fit` without a behavioral-anchor table immediately following
A2. **Anchors present on every dimension.** Each rubric dimension must have a 1-5 anchor table. Anchors prevent the LLM from scoring on vibes. Halt if any dimension has free-text anchors only.
A3. **Disqualifier list is short and mechanical.** Disqualifiers must be deterministic facts (do-not-poach list, location mismatch, staleness). Halt if a disqualifier requires judgment (e.g. "not a culture fit", "seems junior").
A4. **Comp band is recruiter-internal.** The skill's output must not quote a comp number to the candidate. Outreach templates reference the band as "competitive range disclosed on screen". Halt if the rubric includes a "send comp in outreach" instruction.
## B. Pool checks (run after deterministic pre-filter, before LLM ranking)
B1. **Pool size sanity.** If post-filter pool < 10, the skill warns the recruiter that scoring on a tiny pool is meaningless and asks whether to broaden the query. If pool > 200, the skill caps at 200 and notes the truncation in the audit log.
B2. **Geographic spread sanity.** If 100% of post-filter candidates are from one city for a remote-eligible role, the skill warns that the query likely has an over-narrow location filter. Recruiter confirms or broadens.
B3. **Tenure-pattern sanity.** If 100% of candidates worked at the same employer, the skill warns that the query is functioning as a target-list poach rather than open sourcing. Recruiter confirms or broadens.
## C. Output checks (run before writing shortlist)
C1. **Every score above 1 has an evidence string.** Scores without a cited evidence string from the candidate's profile are reset to 1. The skill notes the reset count in the audit log.
C2. **No protected attribute appears in the shortlist or in any outreach draft.** Skill greps the output for the A1 patterns before writing. Hit → halt.
C3. **Skipped candidates are listed, not erased.** The shortlist's "Skipped" table includes every candidate the deterministic filters removed, with the reason. This is what makes the run auditable.
## D. Run-level checks
D1. **Audit log written.** A run is not complete until the JSONL line is appended to `audit/<YYYY-MM>.jsonl`. No PII in this line.
D2. **Human-review gate enforced.** No `send`, `contact`, or `mark_contacted` API call exists in this skill's code path. If you are asked to add one, refuse and surface the request to the user.
## NYC LL 144 / EU AI Act note
This skill is designed to fall *outside* the bias-audit threshold by:
- Producing a ranked list, not an automated decision (no auto-reject).
- Stopping at a human-review gate before any candidate is contacted.
- Logging rubric SHA-256 + pool sizes per run for reproducibility.
If your deployment changes any of those properties (e.g. you wire a "send" action into the loop), you have crossed into automated decision-making and a bias audit is required before production use. NYC LL 144 requires the audit within one year before use; EU AI Act classifies this as Annex III high-risk under Art. 6.
# Shortlist output format
> The candidate-sourcing skill writes `shortlist.md` per the structure
> below. The format is fixed because downstream consumers (the recruiter,
> a hiring manager, an audit reviewer) need predictable columns. Do not
> reformat without updating the skill's output check.
## File: `shortlist.md`
```markdown
# Sourcing shortlist — {Role title}
Generated: {ISO 8601 timestamp}
Channel: {juicebox | hireez | linkedin_recruiter}
Pool: {pre_filter} → {post_filter} → top {n}
Rubric SHA-256: {first 12 chars}
Run ID: {uuid}
## Top {n}
| # | Name | Current role | Current company | Skill | Level | Pattern | Response | Aggregate | Source |
|---|---|---|---|---|---|---|---|---|---|
| 1 | {Name} | {Role} | {Company} | 5 | 5 | 4 | 4 | 18 | {URL} |
## Evidence — top 5
For each of the top 5, cite the specific profile string for every score
above 1. No citation → score reset to 1 (see fairness checklist C1).
### 1. {Name} (aggregate {N})
- **Skill ({score})**: "{verbatim profile excerpt}" — {profile section}.
- **Level ({score})**: "{excerpt}" — {section}.
- **Pattern ({score})**: "{employer sequence}" — {explanation against rubric}.
- **Response ({score})**: profile updated {date}, "{tag if any}".
### 2. {Name} (aggregate {N})
...
## Skipped — surfaced for review (NOT auto-rejected)
| Name | Reason | Source |
|---|---|---|
| {Name} | "current company on do-not-poach list ({customer})" | {URL} |
| {Name} | "stated location {city} outside role policy {policy}" | {URL} |
| {Name} | "profile last updated {date}, staleness > 18mo" | {URL} |
## Suggested talk-track per top candidate
The recruiter uses these as talk-track scaffolding for the first
screening call. They are NOT scripts.
### 1. {Name}
- **Open with**: their {recent role / talk / OSS contribution} — specific
reference, not a generic compliment.
- **Likely motivation hypothesis**: {evidence-based, e.g. "third fintech
role in a row, may be looking for a non-fintech reset; ask"}.
- **Hesitation to surface**: {e.g. "current company is well-funded; ask
what would have to be true for them to consider a move"}.
### 2. {Name}
...
## Outreach drafts
Drafts written to `outreach/{candidate-id}.md`, one file per candidate.
The recruiter reviews, edits, and sends through the ATS or sourcing
tool's outbox. The skill does not contact candidates.
- `outreach/{id-1}.md`
- `outreach/{id-2}.md`
- ...
```
## File: `outreach/<candidate-id>.md`
```markdown
# Outreach draft — {Name}
Channel: {LinkedIn InMail | email | Juicebox sequence}
Subject: {≤60 chars, references a specific signal from the profile}
---
Hi {first name},
{One sentence referencing a specific, recent thing from their profile —
the {recent role / talk / project / post}. Not a flattery line.}
I'm hiring a {role title} at {company}. The reason I reached out is
{specific connection between their background and the role — cite the
profile signal}. The role's {one specific differentiator that would
matter to someone with this background}.
If you're open to a 15-minute conversation, I'm happy to share more. The
comp range will be disclosed on screen if we get to that step.
{Recruiter name}
---
## Recruiter-only metadata (strip before sending)
- Aggregate score: {N}
- Top evidence string: "{excerpt}"
- Source URL: {URL}
- Run ID: {uuid}
- Reviewed by recruiter: [ ]
- Sent: [ ]
```
## Why these fields are non-negotiable
- **`Source` URL on every row** — required for the recruiter to spot-check the LLM's evidence claims against the actual profile.
- **`Pool: pre → post → top N`** — surfaces how many candidates were filtered out deterministically vs. by the LLM. Big LLM-side cuts on a small post-filter pool is a signal of overfitting to rubric noise.
- **`Rubric SHA-256`** — proves which rubric was used on this run (NYC LL 144 audit defense + EU AI Act traceability).
- **`Skipped` table** — candidates filtered out are listed with reasons, not erased. Erasing them turns the workflow into automated rejection.
- **Recruiter-only metadata in outreach** — stripped before sending; its presence in the draft is what reminds the recruiter the message is a draft, not a finished product.
