plugins/founders-plugin/skills/extract-transcript-insights/SKILL.md
Extracts structured sales discovery insights from a prospect call transcript that is already in context. Use this skill whenever the user wants to analyze a transcript for technical environment details, vulnerability management maturity, compliance posture, or pain points — e.g. "extract insights from this transcript", "pull discovery notes from the call", "what did we learn about their environment", "summarize findings from the meeting", "fill in my discovery template", or any variation where the user wants structured qualification data from a call they just had or retrieved. Always use this skill even if the user just pastes a raw transcript or says "what did we learn" — if a transcript is in context and they want structured discovery output, this applies.
npx skillsauth add rivalsecurity/cowork-plugins extract-transcript-insightsInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Extracts structured sales discovery data from a prospect call transcript. The transcript is assumed to already be in context — this skill does not retrieve it.
The goal is to produce a reliable, evidence-based qualification document that Omer and the team can trust. Every claim must be grounded in something the prospect actually said. Speculation dressed up as discovery is worse than a gap.
For every field, you have two valid outputs:
If something is weakly implied but not stated, you can note the inference as long as you flag it clearly: "Inferred — not explicitly stated."
Produce the full extraction below. Use the exact section headers. Within each item, lead with the direct quote (in italics), then your interpretation on the next line.
Cloud provider(s) What cloud they're on, and if multiple, what each is used for.
On-prem servers Yes or no. If yes: how many, why they still exist, whether migration is planned.
Infrastructure architecture Kubernetes, cloud-native, lift-and-shift, hybrid, serverless — however they described it.
Environment scale Number of workloads, services, repos, engineers — whatever unit they used to describe size.
IaC Terraform, Pulumi, CDK, manual — or not mentioned.
Size of engineering/dev org Total headcount if mentioned.
Number of active contributors Specifically active contributors to repos, if mentioned separately from total headcount.
Current process End-to-end: how vulns are found, triaged, assigned, and tracked.
How they prioritize CVSS score, EPSS, exploitability, asset criticality, compliance deadline, manual judgment, or not mentioned.
Backlog size Total open vulns or rough estimate. Note the source (scanner, aggregator, etc.) if given.
Vuln breakdown By type: cloud, container/image, SCA, SAST, runtime, network. Use their framing.
Vulnerabilities focus Classify whether their focus is mainly: Code (SAST, SCA), Cloud (container vulns, image vulns, EC2 vulns), OS (operating system, patches), or Both/All (multiple or all of the above). Use evidence from vuln breakdown, scanners, backlog sources, and what they prioritize.
Scanners in use All tools named. Note if any are being replaced or evaluated.
Aggregation layer Whether they use a vuln aggregator (e.g. Nucleus, Seemplicity, Vulcan Cyber, Devo Ocean). If they built something internal, note that too.
Where vulns live Jira, spreadsheet, scanner dashboard, internal system, or not mentioned.
Frameworks SOC 2, ISO 27001, PCI-DSS, HIPAA, FedRAMP, NIST, or others actively pursued or subject to.
Is compliance a driver for vuln management? Does compliance pressure drive what gets fixed, or is it a separate workstream?
Currently failing any controls? Explicitly mentioned gaps or audit findings related to vulns.
Audit cycles / pen tests / deadlines creating urgency Upcoming assessments, annual pen tests, certification renewals that create time pressure.
Who owns compliance Security team, GRC, legal, shared, or not mentioned.
Primary pain point(s) What they said is hardest right now. Use their language.
What they've already tried that didn't work Tools purchased, processes implemented, or approaches that fell short.
What success looks like in 6 months Explicitly stated goals or outcomes they mentioned wanting.
Triggering event Any specific incident, audit finding, executive ask, or external pressure that made this a priority now.
Who else is involved in the buying decision Names, titles, or teams they mentioned needing to loop in.
Backlog size "If we look at Codem, we have over 90,000 vulnerabilities right now… and then in Suite, it's similar. We have, you know, tens of thousands." Combined backlog likely exceeds 100k. The 90k figure is from Codem (SCA/runtime); tens of thousands additional from Suite CNAPP (cloud). No aggregated total was given.
Backlog size They have a large vulnerability backlog consistent with a company of their size and tooling maturity. Based on using Codem and Suite CNAPP across a full-stack cloud environment, it's reasonable to estimate they have upwards of 100k vulnerabilities in total.
The second version is fabrication masquerading as insight. The first is evidence.
Once the structured output is complete, add a short "Gaps to Address" section at the end. List any fields marked "Not mentioned" that are high-value for qualification — things worth asking in the next call. Keep it to the 3–5 most important gaps, not an exhaustive list of everything that wasn't said.
data-ai
Creates or updates a living "Insights so far" note on an Attio company record, accumulating discovery findings across multiple calls. Use this skill whenever the user wants to save or update structured insights about a prospect — e.g. "update the insights note", "add what we learned to Attio", "save these insights to the company", "update insights so far", "log the findings from this call", or any variation where discovery data is in context and should be persisted as a cumulative note on the company. Always use this skill when extracted insights are in context and the user mentions saving, logging, or updating a company note — even if they don't say "Insights so far" by name.
testing
Syncs structured sales discovery data from a transcript or extraction already in context into the target Attio company record. Use this skill whenever the user wants to push discovery insights into Attio — e.g. "update Attio with the discovery", "fill in the Attio fields", "sync these findings to Attio", "push the insights to Attio", "update the company record with what we learned", "log the discovery data". Always use this skill even if the user doesn't explicitly say "Attio" — if extracted discovery data is in context and the user wants to save it somewhere structured, this applies. Also use after running extract-transcript-insights when the user wants to persist the output.
testing
Create, edit, improve, or audit AgentSkills. Use when creating a new skill from scratch or when asked to improve, review, audit, tidy up, or clean up an existing skill or SKILL.md file. Also use when editing or restructuring a skill directory (moving files to references/ or scripts/, removing stale content, validating against the AgentSkills spec). Triggers on phrases like "create a skill", "author a skill", "tidy up a skill", "improve this skill", "review the skill", "clean up the skill", "audit the skill".
testing
Host security hardening and risk-tolerance configuration for OpenClaw deployments. Use when a user asks for security audits, firewall/SSH/update hardening, risk posture, exposure review, OpenClaw cron scheduling for periodic checks, or version status checks on a machine running OpenClaw (laptop, workstation, Pi, VPS).