All skills
deanpeters avatar

/intelligence-collection-disciplines

@ba3fc8c

Run competitive research like an intelligence agency: eight collection disciplines (OSINT to MASINT), signal-to-inference chains, and fusion. Use when one-source research isn't enough.

Use this Skill: https://skilld.dev/gh/deanpeters/product-manager-skills/intelligence-collection-disciplines

This session only. Nothing lands on disk.

template.md

≈588 tokens on demand. Your agent reads this file only when SKILL.md points to it.

Collection Plan Template

Copy this to plan a collection effort before running any discipline. The plan turns the compendium from a course into an operation: fill the six instantiation variables, pick the two or three disciplines that feed your decision (never all eight), and set the cadence to how fast each evidence channel actually changes. If [DECISION] is blank, stop — research without a decision is a hobby.

Template

# Collection Plan: [TARGET] / [CAPABILITY]

## Instantiation
[TARGET]      = [the competitor, partner, or acquirer]
[MARKET]      = [category + NAICS/SIC/NACE codes or nearest equivalent]
[GEOGRAPHY]   = [regions in scope, at COUNTRY level]
[BUYER]       = [who signs the check]
[CAPABILITY]  = [the strategic move you suspect]
[DECISION]    = [what this research will change — REQUIRED]

## Discipline Selection
_Pick from the artifact-mapping table — run the disciplines that feed the decision, not all eight._

| Discipline | Running? | Why (which signal chains) | Key sources (free first) |
|---|---|---|---|
| OSINT | | | |
| FININT | | | |
| GEOINT/DEMOINT | | | |
| TECHINT | | | |
| HUMINT | | | |
| SIGINT | | | |
| MASINT | | | |

## Cadence
- Weekly (SIGINT sweep, ~30 min): [what gets checked]
- Monthly (OSINT + HUMINT digest): [what gets checked]
- Quarterly (FININT + TECHINT deep pass): [what gets checked]
- Event-driven (react within 48h): [which tripwires]

## Fusion Table (fill as signals land)
| Discipline | Signal found | Label | Source URL + date |
|---|---|---|---|
| | | | |

**Fusion verdict:** ____ disciplines, one story → [watch item / working hypothesis / actionable] → recommended response: ____

## Provenance & Guardrails
- Every signal archived at collection time (portals restructure; files get replaced)
- All sources published, filed, posted, or publicly observable — the stage test applies to every method

Quality checks before you call it done

  • [DECISION] is filled in and specific — it selects the disciplines and defines "enough"
  • Two or three disciplines are marked Running, with chain-level reasons — eight checkmarks means no plan
  • The fusion verdict uses the stacking rule (1 = watch, 2 = hypothesis, 3+ = actionable, conflict = dig)
  • Announcements in the fusion table are labeled intent until FININT/MASINT/HUMINT corroborate

Source: SKILL.md on GitHub

1 warning1mo3 checks · Risk SAFE
  • Gen Agent Trust Hub1mo

    The skill provides a framework for competitive intelligence research. While the methodology is sound, it inherently involves processing data from untrusted external sources, presenting a low risk for indirect prompt injection.

  • Socket1mo

    No alerts

  • Snyk1mo

    Risk: MEDIUM · 1 issue

Signed by skilld at ba3fc8c. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub last month.

Activeupdated 2 months ago
type
component
theme
market-intelligence
estimated_time
reference skill; a single-discipline pass takes 30-60 min
Other metadata
intent
Give product managers the intelligence community's collection playbook: eight independent disciplines, each with free and paid sources, signal-to-inference chains, and the PM artifact it feeds — then fuse them with confidence stacking so three weak signals become one strong conclusion.
best_for
[
  "Choosing which collection channels to run for a competitive or market question",
  "Turning raw public signals (patents, job posts, filings, web diffs) into defensible inferences",
  "Cross-validating a suspected competitor move across independent evidence channels"
]
scenarios
[
  "I think a competitor is building a platform play — how do I confirm it before their launch?",
  "My TAM slide got shredded — where do I find data that survives scrutiny?"
]

README badge

README badge for deanpeters/product-manager-skills/intelligence-collection-disciplines