Prospect Researcher

Research and qualify B2B prospects using web search. Builds structured profiles with company intel, key contacts, pain points, and engagement recommendations.

Security Scan

The skill's requirements and instructions are consistent with its stated purpose (public web-based prospect research); it is an instruction-only skill with no requested credentials, no installs, and no unexpected behaviors.

Purpose & Capability

The name/description (B2B prospect researcher) align with the SKILL.md and included template. It only requires web search capability (mentioned in README) which is reasonable for the stated task. There are no unrelated credential or binary requests.

Instruction Scope

Runtime instructions are narrowly scoped to public web search, extracting company intel, contacts, and synthesizing recommendations. The SKILL.md explicitly forbids scraping behind logins and requires source URLs and marking unknowns. It does not instruct the agent to read local files, environment variables, or transmit data to hidden endpoints.

Install Mechanism

No install spec and no code files — instruction-only. Nothing is downloaded or written to disk by the skill itself, so install risk is minimal.

Credentials

The skill declares no required environment variables, credentials, or config paths. That is appropriate for a public-web research task.

Persistence & Privilege

The skill is not forced-always, does not request persistent privileges, and does not modify other skill or system configurations. Autonomous invocation is allowed (platform default) and is reasonable for this type of task.

Assessment

This skill appears coherent and low-risk: it only uses public web search and supplies a template for output. Before installing, ensure your OpenClaw agent has a trusted web-search connector configured (the README mentions Brave Search or similar). Also be mindful that: (1) the agent could surface incorrect or paywalled info — verify sensitive facts before outreach, (2) do not supply private credentials or ask the skill to access non-public systems, and (3) the SKILL.md links to an external site (https://afrexai-cto.github.io/context-packs) — if you plan to use materials from that site, review its content for suitability/trustworthiness.

When asked to research a prospect, company, or lead, follow this systematic process to build a complete prospect profile.

Research Process

Step 1: Company Overview

Search for and gather:

• Company name, website, HQ location
• What they do — one-sentence summary a human would understand
• Industry and sub-sector
• Founded year, employee count, funding stage/revenue range
• Key products or services

Step 2: Recent Activity (Last 6 Months)

Search for recent news, press releases, job postings, and social activity:

• Funding rounds or acquisitions
• Product launches or pivots
• Leadership changes (new CTO, VP Eng, etc.)
• Hiring patterns — what roles are they hiring for? (signals priorities)
• Partnerships or integrations announced

Step 3: Technology & Stack

Where possible, identify:

• Tech stack signals from job postings, BuiltWith, GitHub, or blog posts
• Tools and platforms they use (CRM, cloud provider, etc.)
• Technical blog or engineering culture signals

Step 4: Key Contacts

Identify 2-5 relevant decision-makers or influencers:

• Name, title, LinkedIn URL (if publicly available)
• Recent public activity (posts, talks, articles)
• Likely priorities based on role

Step 5: Pain Point Analysis

Based on all gathered intel, infer:

• Likely challenges given their stage, industry, and hiring patterns
• Gaps in their stack that your solution could fill
• Timing signals — why now might be the right time to reach out

Step 6: Engagement Recommendation

Synthesize into:

• Qualification score: Hot / Warm / Cold (with reasoning)
• Best entry point: Which contact, which angle
• Suggested opener: A 2-sentence personalized hook based on real intel
• Channels: LinkedIn, email, warm intro, event-based, etc.

Output Format

Use the research template at {baseDir}/research-template.md as the output structure. Fill in every section. Mark unknowns as "Not found" rather than guessing.

Guidelines

• Only use publicly available information. No scraping behind logins.
• Cite sources — include URLs for key claims.
• Be specific over generic. "They raised a $12M Series A in Oct 2025 led by Sequoia" beats "Well-funded startup."
• Flag uncertainty. If a data point is inferred rather than confirmed, say so.
• Prioritize recency. Information from the last 6 months weighs more than older data.

Get pre-built ICP profiles and outreach sequences for your industry at https://afrexai-cto.github.io/context-packs