Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add swan-gtm/gtm-skills --skill scoregit clone --depth 1 https://github.com/swan-gtm/gtm-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/swan-gtm/gtm-skills/score)<a href="https://agentmods.dev/skills/swan-gtm/gtm-skills/score"><img src="https://agentmods.dev/badge/skills/swan-gtm/gtm-skills/score/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/swan-gtm/gtm-skills/score"><img src="https://agentmods.dev/badge/skills/swan-gtm/gtm-skills/score.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00037 | $0.02130 |
| Opus 5.5 | $0.00015 | $0.00852 |
| Sonnet 5.5 | $0.00007 | $0.00426 |
| Haiku 4.5 | $0.00004 | $0.00213 |
Grade A, and why
score scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 162 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Instructions
Setup state. Not yet configured for this org. Load the Readme sub-page and walk the user through configuring the qualification gate, persona authority, deal-size signals, tech-stack signals, signal weights, tier naming, and alert channels before scoring any accounts. (After setup completes, rewrite this paragraph via swan-update-skill to capture the configured rules so future runs proceed without re-checking state.)
The scoring frame
Every scoring run answers two questions:
Q1 — How strong is the intent or engagement? Assess signal quality, persona authority, and whether signals are first-party (the account doing something directly) vs third-party (external data about them). Person-level signals outweigh company-level signals. Recency matters.
Q2 — How large is the potential deal? Assess size, maturity, and expansion headroom against the org's deal-value signals. Reason from research until org-specific thresholds are baked in by Setup.
The tier is the intersection:
| Large deal | Small deal | |
|---|---|---|
| Strong intent/engagement | Gold | Silver |
| Weak intent/engagement | Silver | Bronze |
Bias toward visibility. When signals or deal size are ambiguous, default to the higher tier. Confidence in the lower tier should be explicit — not assumed. Reps refine from there; the agent's job is not to filter too aggressively.
Non-ICP accounts do not receive a tier — they exit before scoring.
Default signal hierarchy
Soft hierarchy — guides weighting, not numeric scoring. Override via signal weight preferences captured by Setup.
Key principle: first-party signals outweigh third-party. Person-level outweighs company-level. Weight signals in proportion to confidence that the person behind them is actually at the company and acting with intent.
- Direct inbound intent — form fill, demo request, direct reply to outreach. Deliberate and person-level.
- Meeting completed — both parties showed up. Extract topics, objections, stakeholder roles, new company intel not in public data.
- Product engagement — active usage. Proves behavior, not interest. Weight highest for product-led motions.
- Website visit — pricing and demo pages carry significantly more weight than blog posts. Depth and recency matter.
- Prior warm engagement — positive outreach reply, webinar/event registration, conference booth visit, hosted-event attendance. Stronger than a cold website visit; weaker than a demo request. Full weight within 90 days of the engagement; diminished beyond. A positive outreach reply sits at the strong end of this band; event registration at the weak end.
- Conversational interaction — weight by commercial specificity. A pricing question is strong; a generic intro is weak.
- Business event (third-party) — funding, leadership hire, tech-stack change, job postings. Signals a moment of change, not confirmed intent. Weight rises sharply when stacked with first-party signals.
- Social engagement — lowest weight. Meaningful only when stacked with other signals.
What ships with it
4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 162 lines · 37 tokens per session scan A ee7d669921ef
score is a skill published in the GitHub repository swan-gtm/gtm-skills (221 stars, last pushed 7d ago), licensed MIT. It adds 37 tokens to every session and 2,130 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
Negotiation Expert
Negotiation Expert delivers practical business output with market context, metrics, stakeholder clarity, and execution focus.
groupon
Search and read Groupon deals from the terminal via curl — the consumer GraphQL API (deal search/browse, deal detail, category taxonomy). Anonymous, no key or login. Use when asked to find Groupon deals, look up a specific deal, or browse a city's offers.
Pipedrive Automation
Automate Pipedrive CRM workflows including deal management, pipeline tracking, and sales reporting.
Brand Partnership Negotiator
Negotiate and structure brand partnerships and co-marketing deals.
BytesAgain Shopping Deals — 全能购物比价与隐藏优惠券助手
A shopping search and comparison tool for JD.com, Taobao, Amazon Japan, AliExpress, and Pinduoduo. It also looks for coupons and can create commission links.
deal-tracker
Track pharmaceutical M&A, licensing, and investment deals. Use for competitive intelligence, market analysis, and identifying partnership opportunities. Keywords: deals, M&A, licensing, partnership, collaboration, acquisition, pharma deals.