Analyzing Competitors

Analyzing Competitors

Run a Black Hat competitive assessment and record it as a living subgraph of the pursuit wiki: one entity page per competitor, each modeling how that competitor will bid. The core technique is perspective — bid the job as each competitor would. The wiki payoff is that the assessment does not sit in one flat document: each competitor page links to the customer hot buttons it threatens, the team's exposures it creates, and later the win themes built to counter it.

When to use this skill

Use this skill during capture, once a pursuit wiki exists (created by qualifying-opportunities), to assess the competitive field. It enriches any low-confidence competitors/ stubs the bid-decision left behind, and it feeds price-to-win and win-strategy work.

Architecture

This skill is wiki-native. Read [docs/architecture.md](../../docs/architecture.md) and the [conventions](../../conventions/) first. It reads sources and entities, maintains competitors/ and risks/ entity pages, and renders the competitive-assessment.md view. Relevant domain-wiki pages: [[discriminator]], [[best-value-tradeoff]], [[lpta]].

Inputs

Required: an existing pursuit wiki. Read pursuit.md (note the posture — it frames the whole assessment), the customer/ entities, the decisions/bid-decision.md entity, and any competitors/ stubs.

Private company context (recommended). If pursuit.md carries a company_context_path:, read these company-context folders before intake:

• partners/: competitors that are also your partners on other deals, or

former partners, or candidate teammates here. The relationship history is load-bearing.

• debrief-lessons/: what prior debriefs told you about how these

competitors actually bid against you.

• customers//: your prior interactions with this customer

that reveal what the customer thinks of each competitor.

See [../../docs/customize-for-your-company.md](../../docs/customize-for-your-company.md).

Before any competitor research, read references/competitive-intelligence-ethics.md. Only properly sourced information may enter a competitor entity page.

Intake

Ask these as a numbered list. Skip what existing entities answer.

1. The field. Who are the known and suspected bidders? Is there an

incumbent, and who is it? (The set-aside and vehicle in pursuit.md bound the field.)

1. What is known per competitor. For each likely bidder — size, relevant

past performance, customer relationship, typical technical approach, typical price posture, recent wins and losses. What is the source of each claim?

1. The incumbent's standing. If the incumbent is a competitor, how is the

customer relationship — strong, strained, fatigued? Known performance problems?

1. The team's position. Briefly, where is the team strong and weak relative

to this field? (Used to assess exposure, not to set strategy.)

1. Sources. What public or properly obtained sources are available?

Where the team does not know something, that is an intelligence gap — record it on the competitor entity with confidence: low; do not invent it.

Workflow

Black Hat Competitive Assessment:
- [ ] Step 1: Open the pursuit wiki; read company-context if configured; read CI ethics
- [ ] Step 2: Complete intake; confirm the competitive field
- [ ] Step 3: Create or enrich a competitor entity per bidder
- [ ] Step 4: Bid the job as each competitor (the Black Hat)
- [ ] Step 5: Record ghosting opportunities and the team's exposures
- [ ] Step 6: Emit claims/ and source-register/ entities for the competitor reads
- [ ] Step 7: Render the competitive-assessment view; validate

Step 1: Open and read. Read pursuit.md, the customer entities, the bid decision, the competitor stubs, and the CI ethics reference. If pursuit.md carries a company_context_path, read the relevant partners/, debrief-lessons/, and customers/ company-context pages — they are first-rung sources for prior-relationship facts. Cite the company-context pages on the resulting competitor and source entities.

Step 2: Confirm the field. List the credible bidders; distinguish likely from possible; treat the incumbent as its own case.

Step 3: Competitor entities. For each bidder, create or enrich entities/competitors/.md per references/black-hat-method.md — the ## Profile section (qualifications, past performance, relationship/incumbency, typical approach, typical price posture, capacity, recent wins/losses). Tag every claim's source and set the page confidence to the weakest claim it relies on. Mark each item known or assessed.

Step 4: Bid as them. For each competitor entity, write the ## Assessed offer section by switching perspective: would they bid, their likely solution, their likely win themes, their genuine strengths and vulnerabilities on this bid, and their likely price posture. Give each competitor their best plausible bid — no strawmen.

Step 5: Ghosting and exposures. On each competitor page, write a ## Ghosting opportunity section — a real, evidence-based competitor weakness, framed as a positive, proven statement of the team's strength (no competitor named in proposal text; the entity page may name them). Then create one entities/risks/.md page per exposure the team carries that a competitor will exploit, linked from the competitor entity — these feed developing-win-strategy.

Step 6: Claims and source-register. Emit the evidence graph:

• entities/claims/: write one claim entity per assertable competitor

fact ("Acme Corp is the incumbent on the predecessor contract," "Acme's typical labor mix runs 30% senior / 70% mid-level," "Acme has current TS/SCI bench across the region"). Each claim has the claim text, the owner who can defend it, the assertion type (known fact / assessed / hypothesis), and a status.

• entities/source-register/: write one source entity per external

source the claims rest on: SAM.gov / USAspending records, Acme's public website and press releases, current public job postings, prior public awards, and any cited company-context page. Tag the source's rung on the [source-policy.md](../../knowledge/source-policy.md) authority ladder.

For competitor-proprietary or source-selection-sensitive material that the team must not use, do not create a claim or source entity. Quarantine it per [competitive-intelligence-ethics.md](references/competitive-intelligence-ethics.md).

Step 7: Render and validate. Render views/competitive-assessment.md from the competitor, risk, claims, and source-register entities, including the comparison matrix and an honest most-probable-winner read. Run the validation checklist in black-hat-method.md. Update pursuit.md.

Output

A competitors/ subgraph (one entity per bidder), risks/ entities for the team's exposures, supporting claims/ and source-register/ entities, and the competitive-assessment.md view.

Tell the user: the likely field, the most-probable winner and why, the strongest ghosting opportunities, the team's most dangerous exposures, and the biggest intelligence gaps.

References

• references/black-hat-method.md: competitor entity sections, the "bid as them" technique, the comparison matrix, ghosting, exposures, and the validation checklist.
• references/competitive-intelligence-ethics.md: which competitor information may be used, and how the sensitivity tag quarantines what may not.

Guardrails

• Properly sourced information only. Never use, seek, or speculate from a

competitor's proprietary information or any source-selection-sensitive material. Quarantine such material with the sensitivity tag; do not fold it into an entity. See the ethics reference.

• No strawmen. Give each competitor their strongest plausible bid.
• Mark assessed vs. known. A predicted competitor offer is an assessment —

set confidence honestly; never state a prediction as fact.

• Be as honest about us as about them. The team's exposures get the same

candor as competitors' weaknesses, recorded as real risks/ entities.

• Maintain the graph, render the view. The competitor, risk, claim, and

source-register entities are the source of truth; the assessment view is a rendered read-out.

• Company-context stays private. When the skill draws on a configured

company_context_path, it references the company-context pages by path; it does not copy their content into the pursuit wiki. Resulting pursuit entities carry sensitivity: company-proprietary where the underlying fact came from company-context. See [../../docs/customize-for-your-company.md](../../docs/customize-for-your-company.md).

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: danielkinneyspears
• Source: danielkinneyspears/govcon-pursuit-brain
• License: Apache-2.0

Install and usage instructions live in the source repository linked above.