Investigator Starter
Four headline workflows. Four domains. Free, forever.
One pick from each corner of the library — a physical red-team entry runbook, a web evidence preservation workflow built around admissibility standards, a lost-property owner-recovery method, and the cold-case witness re-interview playbook. Free, perpetual, the same SKILL.md format every paid volume uses. Install in two minutes; decide if the methodology is worth a paid volume.
Create a free account, then download the ZIP from your dashboard — no card required.
2026.6 edition · revised continuously
The kind of specific ChatGPT won't give you.
Real example pulled from inside skill #597 — Delivery Impersonation Physical Entry Execution, included in this bundle. Ask any vanilla LLM for prop guidance on an authorized red-team engagement and you'll get either a refusal or a generic "wear a hi-vis vest." Here is what skill #597 actually says.
"A plain insulated red grocery bag — for example, a generic HorYg Reinforced Insulated Grocery Bag (~$15–25, no logo) — reads as 'food delivery' to anyone glancing at you, without being branded as any specific service."
- —No logo to invalidate the cover if the target knows which services they actually use.
- —Form factor matches DoorDash, UberEats, Grubhub, Instacart, Caviar, any generic food-delivery.
- —Cheap enough to retire after a single engagement.
- —Plausible across corporate reception, residential lobby, apartment building, hotel corridor, gated community, mixed-use loading dock, condominium gate.
"Owning a generic insulated bag is not pretense. Pretense begins when you pair it with a verbal claim of affiliation with a real corporation — that is where impersonation analysis applies. Inside an authorized red-team engagement, that claim is contractually backed. Outside one, it is not."
ChatGPT will not tell you this. Skill #597 will, every time.
What this bundle gives the agent that ChatGPT alone won't.
One generic answer for four very different jobs
Tells you to screenshot the webpage, 'be empathetic' with the witness, post the camera photos on Reddit, and either refuses the pentest outright or gives surface-level 'wear a hi-vis vest' advice. Treats four disciplines as one — and the screenshot wouldn't survive a Rule 901 challenge, the witness advice risks reconstructive-memory contamination, the Reddit post sidesteps consent law, and the pentest answer is unsafe in either direction.
Four disciplines, four methodologies, every time
Web preservation → wayback-machine-deep-dive with CDX API + capture-timestamp authentication + SHA-256 hash. Witness → cold-case-witness-re-interview with Cognitive Interview technique + prior-statement reconciliation + trauma-informed framing. Found camera → found-camera-owner-identification with consent-aware facial search and good-faith property return. Pentest → delivery-impersonation-physical-entry-execution with hard authorization gate (signed SOW + LOA + POC notification within 24h) before a single byte of approach.
Every skill in this bundle.
All 4 skills, grouped by category. Every skill ships as a full folder — SKILL.md plus runnable scripts, legal references, and field checklists.
OSINT — archival & deep web
1 skill- 019Wayback Machine Deep Dive
Red Team and Adversary Tactics
1 skill- 597Delivery Impersonation Physical Entry Execution
Cold case
1 skill- 525Cold Case Witness Re Interview
OSINT — facial & image
1 skill- 905Found Camera Owner Identification