一键导入
gtm-context-os
gtm-context-os 收录了来自 thesnappingdog 的 15 个 skills,并提供仓库级职业覆盖和站内 skill 详情页。
这个仓库中的 skills
Dashboard of current GTM state across all active modules. Use when user asks about status, progress, or what to do next.
Upgrade this GTM Context OS instance to the latest template developments. Fetches the latest template, reconciles what's new against THIS instance's structure, and applies approved changes on a review branch — pausing before anything destructive. Operator-initiated.
Guided first-use setup. Walk through filling context.md and ingesting the first sales call. Use when the repo is empty or context.md is unfilled.
Smoke test before merging dev to main. Bootstraps a clean worktree, runs evals, checks instruction consistency — all automated via sub-agents.
First thing to run. Checks repo state and recommends what to do next. Use when operator says 'what do I do', 'help me get started', 'where do I begin', or just seems lost.
Audit the health of THIS instance's lived-in state — structural completeness, evidence chains, index integrity, output hygiene, drift, and context.md consistency (conflicts + cached-vs-source drift). Reports findings and fixes safe drift; flags judgment calls. Runs inside Claude Code, no API key.
Process raw documents into structured, attributed knowledge. Drop docs into _intake/, run this skill to dissect them into the right places. Incremental — run as many times as needed.
Connect a specific tool — CRM, call recorder, enrichment, sequencing. Scaffolds scripts, MCP config, and integration docs. Use when operator says 'connect HubSpot' or 'I need to pull transcripts from Gong'.
Run the eval suite against the current repo state. Tests whether AGENTS.md instructions produce correct behavior. No API key needed — runs inside Claude Code.
Draft an outbound sequence grounded in demand data. Use when messaging angles are defined and ready for copy.
Search PULL analyses by criteria to find evidence for segments. Use when searching for demand signals, conversion patterns, or buyer evidence.
Generate a detailed handover message for continuing work in a new session. Use at end of session or when context window is getting long.
Fastest way to get started — give a company website URL and agents crawl it to populate context.md. Use when operator wants to set up fast without answering 20 questions.
Check and set up base technical environment — Python, uv, .env, .mcp.json. Use when repo is freshly cloned or operator hits a missing-tool error.
Match a segment to messaging angles based on PULL evidence. Use after segmentation when developing outreach angles.