Skip to content

getaero-io/gtm-eng-skills / engineering/warm-intro-scoring

warm-intro-scoring

Verified

FileKindSize
.gitignoreother121 B
GUIDE.mdmarkdown4.7 KB
LICENSEother1.1 KB
README.mdmarkdown8.0 KB
SKILL.mdskill md327 B
TESTING.mdmarkdown8.9 KB
UPSTREAM.mdmarkdown1.3 KB
assets/company-research.play.tsscript1021 B
assets/customer-db.sqlother2.7 KB
assets/evaluation-example.jsonconfig4.6 KB
assets/example.jsonconfig6.9 KB
assets/portfolio-focus.weights.jsonconfig32 B
assets/review.htmlother7.4 KB
assets/tuning-example.jsonconfig1.2 KB
assets/tuning.htmlother30.1 KB
check_all.pyscript1.4 KB
package-manifest.jsonconfig9.8 KB
plays/README.mdmarkdown18.2 KB
plays/check-report.pyscript1.8 KB
plays/check_all.pyscript1.6 KB
plays/check_parity.pyscript6.0 KB
plays/collect.play.tsscript4.0 KB
plays/collection.tsscript1.9 KB
plays/core.tsscript44.2 KB
plays/evaluation-contract.mdmarkdown3.4 KB
plays/evaluation.tsscript13.8 KB
plays/features.play.tsscript4.7 KB
plays/features.tsscript22.6 KB
plays/generatedtemplate.tsscript38.1 KB
plays/guardrails.tsscript948 B
plays/legacy.tsscript9.7 KB
plays/merge-payloads.pyscript8.6 KB
plays/parity-driver.tsscript995 B
plays/parity-evaluation.pyscript8.5 KB
plays/parity-reports.pyscript2.3 KB
plays/prepare-inputs.pyscript6.5 KB
plays/quality.tsscript3.8 KB
plays/read-sources.play.tsscript3.4 KB
plays/report-cli.tsscript590 B
plays/report.play.tsscript2.9 KB
plays/report.tsscript3.1 KB
plays/reports.tsscript5.7 KB
plays/research.play.tsscript1.6 KB
plays/review-uncertainty.play.tsscript5.5 KB
plays/score.play.tsscript2.3 KB
plays/snapshots.play.tsscript4.2 KB
plays/strict-json.tsscript1.8 KB
plays/sync-template.pyscript849 B
plays/uncertainty.tsscript4.9 KB
plays/validation.play.tsscript3.6 KB
references/adversarial-review.mdmarkdown7.0 KB
references/batch-collection.mdmarkdown5.1 KB
references/company-network.mdmarkdown5.8 KB
references/competitive-review.mdmarkdown8.8 KB
references/evaluation.mdmarkdown8.5 KB
references/evaluator-validation.mdmarkdown3.4 KB
references/evaluator.mdmarkdown7.9 KB
references/external-data.mdmarkdown4.0 KB
references/factor-test-matrix.mdmarkdown4.6 KB
references/factors.mdmarkdown6.1 KB
references/intro-drafts.mdmarkdown1.6 KB
references/pipeline.mdmarkdown7.6 KB
references/readable-output.mdmarkdown5.4 KB
references/research-brief.mdmarkdown2.4 KB
references/sources.mdmarkdown7.5 KB
references/tuning.mdmarkdown14.6 KB
scripts/audit_history.pyscript5.1 KB
scripts/check.pyscript5.1 KB
scripts/compare_versions.pyscript1.9 KB
scripts/evaluate.pyscript22.5 KB
scripts/package-skill.pyscript1.9 KB
scripts/portfolio.pyscript3.7 KB
scripts/quality.pyscript5.3 KB
scripts/recipe-demo.tsscript2.6 KB
scripts/render.pyscript1.6 KB
scripts/score.pyscript9.6 KB
scripts/temporal.pyscript2.2 KB
scripts/test_ask_handoff.pyscript2.1 KB
scripts/test_audit_history.pyscript2.7 KB
scripts/test_handoff_evaluate.pyscript9.1 KB
scripts/test_portfolio.pyscript3.6 KB
scripts/test_quality.pyscript3.1 KB
scripts/test_temporal.pyscript1.3 KB
scripts/test_tuning.pyscript14.1 KB
scripts/test_tuning_browser.cjsscript7.7 KB
scripts/test_version_parity.pyscript1.0 KB
scripts/tuning.pyscript10.9 KB
tests/adversarial.test.tsscript8.2 KB
tests/browser_e2e.cjsscript1.9 KB
tests/cloud_e2e.pyscript6.9 KB
tests/e2e-fixture.jsonconfig3.6 KB
tests/feature_parity.pyscript2.4 KB
tests/features.test.tsscript8.3 KB
tests/test_prepare_inputs.pyscript8.6 KB
tests/uncertainty.test.tsscript2.5 KB
vendor-manifest.jsonconfig748 B
# Warm-intro scoring — engineering recipe

**Start here.** This folder contains Deepline Plays and the offline Python reference for warm-intro scoring. Start with [the Play handoff](plays/README.md) for cloud runs and parity checks. This maintained package lives in `engineering/warm-intro-scoring/`, outside both `drafts/` and the generated `skills/` tree. Start with [TESTING.md](TESTING.md) for the packaged demo. Copy this entire folder into another checkout if needed; The offline Python checks require Python 3.10+. The Play checks also use Bun; cloud runs require the Deepline CLI. No install, credentials, customer database or provider calls are needed for the tests.

```bash
python3 engineering/warm-intro-scoring/plays/check_all.py
```

From inside this folder, run `python3 plays/check_all.py` for the full TypeScript and Python checks. `python3 check_all.py` runs the Python-only subset. It verifies pinned dependencies, runs unit/scoring checks, audits fictional evidence, and generates temporary artifact outputs. Nonzero exit means the handoff is failing.

## Try it

Run from this folder; use fresh output names because scripts refuse overwrite:

```bash
python3 scripts/quality.py assets/example.json --output quality.json
python3 scripts/score.py assets/example.json --output scores.csv
python3 scripts/render.py scores.csv --output paths.html
python3 scripts/evaluate.py assets/evaluation-example.json --output evaluation.json --html evaluation.html
```

Open `paths.html` for the top-three path review or `evaluation.html` for the comparison scorecard. These examples are fictional. The evaluator deliberately reports insufficient evidence.

## What the team is getting

| File                            | Purpose                                                                            |
| ------------------------------- | ---------------------------------------------------------------------------------- |
| `GUIDE.md`                      | Agent workflow: warm connections + optional targets, enrichment, scoring, artifact |
| `scripts/quality.py`            | Pre-scoring quality audit of the existing reviewed-evidence JSON contract          |
| `scripts/score.py`, `render.py` | Deterministic scoring and interactive evidence review                              |
| `scripts/evaluate.py`           | Paired rankings, account-cluster intervals, fixed-window outcome evaluation        |
| `assets/customer-db.sql`        | PostgreSQL reference migration; inspect tenant and current schema before applying  |
| `references/pipeline.md`        | Incremental DB → enrichment → features → scoring deployment contract               |
| `references/evaluator.md`       | Labeling, research-derived challenges, statistical limits                          |
| `vendor-manifest.json`          | Source commit and checksums for the five pinned runtime dependencies               |

The defaults use `vendor/`, independent of the repository's examples and refreshable skills directory. The manifest protects against silent drift. When intentionally refreshing the baseline, review the upstream diff, update the pinned files and hashes together, and rerun checks. The original MIT license is included. Vendored ask code is used only to verify CSV compatibility; `check_all.py` never generates or sends asks.

## Checks added

- Profile identity collisions; conflicting identity values across paths; malformed URLs.
- Missing job history, company or title, reported as unknown with coverage counts.
- Missing/invalid evidence references via the scorer's existing validation.
- Stale receipts and duplicate source receipts, reported as warnings.
- Future employment ends, current-job/end-date contradictions and empty candidate sets.
- Private quality reports and nonzero exit on blocking data errors.

Existing tests cover both-edge holds, latest declines, future features, train/test leakage, ranking math, grouped bootstrap, unknown judgments, mature outcome denominators, horizon boundaries, malicious HTML content and no-overwrite output.

`quality.py` exits **0** for structural pass (warnings can remain), **1** for data-quality failures and **2** for malformed JSON/I/O. Freshness defaults to 180 days relative to the scoring cutoff and is configurable with `--max-age-days`; this is an operational threshold, not a validated relationship-decay weight. Quality reports use IDs and counts, not copied profile prose.

## Before connecting production

Manual provider, database snapshot, feature, scoring, validation, and report Plays are in `plays/`. The separate queue/trigger schema in `references/pipeline.md` is not deployed. The offline checker does not verify that a profile is full or true, query live tables, or implement queue workers. Deployment acceptance must cover transaction rollback, duplicate event replay, lease expiry, out-of-order profile updates, removals, stale artifact invalidation and no extra paid calls after retry. Use source run receipts to reconcile contacts requested/enriched/partial/failed before accepting a batch. Enforce tenant-scoped keys and access controls in the actual database.

Keep raw profiles, connection exports, credentials and real-person artifacts private. Historical data lacks outcome-complete labels; passing this suite does not establish prediction accuracy or business uplift.

## Forward this to the team

> The warm-intro handoff lives in `engineering/warm-intro-scoring/`, outside the refreshable skills tree. Run `python3 engineering/warm-intro-scoring/plays/check_all.py` with Python 3.10+ and Bun. It is offline and credential-free. Start with README.md; the fixtures demonstrate scoring, data-quality reports and evaluation artifacts. Use plays/README.md for the manual cloud workflow and tested parity scope. Automatic triggers, fresh-provider recovery tests, and independently labeled outcomes remain separate deployment gates.

## Adjustable preferences and company networks

```bash
python3 scripts/tuning.py assets/tuning-example.json --output tuning.html
python3 scripts/tuning.py assets/tuning-example.json --weights assets/portfolio-focus.weights.json --output portfolio-focus.html
```

Independent controls cover recorded intros, dated work/school overlap, requester relationships, role/industry, shared investor portfolios, public appearances dated board overlap, work city, industry, professional community and direct investor roles. Export a config to reuse it. The baseline score and holds remain fixed. See `references/tuning.md` for the input contract and `references/company-network.md` for investor discovery/portfolio expansion. Real customer snapshots and research graphs stay outside Git.

## Target-first report

The report shows one target at a time. Select a contact, compare intro paths, then inspect each score breakdown.
Use the weight panel to change ranking preferences. Weight changes do not confirm relationships or permission.
Company research can show source-backed facts from external providers. It does not automatically add score points.
See [external evidence](references/external-data.md) and [clear output rules](references/readable-output.md).

For full account lists, the tuner keeps all candidates, filters by account, and pages the visible paths. Rate a path and add a note, then use **Download feedback** before closing the tab. Reviews include the score and weights used at review time. They do not change permission or readiness. Read the [competitive review](references/competitive-review.md) for gaps found in Centralize, Connect The Dots, and The Swarm.

An optional browser check exercises 91,434 fictional paths, mobile layout, weight changes, feedback exports, and blocked network requests:

```bash
PLAYWRIGHT_MODULE=/path/to/node_modules/playwright node scripts/test_tuning_browser.cjs
```

The core Python checks require no browser dependencies. Real input lists, scores, and feedback stay outside this repository.

## End-to-end and adversarial verification

See [the test report](references/adversarial-review.md) for the cached cloud pipeline, Claude Code findings, corrections, changed-score counts, and remaining prebuilt release gates.
slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack."Stars: 198kUpdated Claude CodeCodexVerifiedInstall
ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt.Stars: 194kUpdated Claude CodeCodexVerifiedInstall