guerrilla2799
UserThe open-source GTM engineering and AI ops layer for Claude Code. 26 skills for the systems underneath the playbooks: data plane, agent building, evals, and martech ops.
Categories
Indexed Skills (25)
agent-guardrails
Use before shipping any agent to another person, or when the user says "guardrails", "the agent did something it shouldn't", "spend limits", "it emailed the wrong person", "safety rails", "what stops it from", "it wrote to the CRM". Defines the refusal set, spend ceilings, mandatory pre-checks, escalation paths, and human gates, then puts the hard ones where persuasion cannot reach them. Writes workspace/agents/<name>/guardrails.md.
agent-observability
Use when instrumenting anything before scaling it, or when the user says "how is the agent performing", "we have no visibility", "instrument this", "logging", "telemetry", "cost per run", "is it getting worse", "correction rate". Defines what to log, what to alert on, and the weekly review that catches degradation before a user does. Writes workspace/evals/telemetry-spec.md.
agent-rollout
Use when shipping an agent to other people, or when the user says "nobody uses it", "roll this out", "how do I distribute this", "adoption", "train the team on it", "package this as a plugin", "version it". Handles packaging, distribution, the first-run experience, accountability, and the adoption metrics that decide whether the build succeeded. Writes workspace/agents/<name>/rollout.md.
agent-scoping
Use before building any AI agent or skill, or when the user says "should we build an agent for this", "can AI do this", "automate this workflow", "where should we use AI", "build me an agent". Decides what to build, what to refuse, and what to leave manual. Produces an agent brief with a success threshold and a kill criterion. Writes workspace/agents/<name>/brief.md. Scopes one agent. Not the org-level AI policy, data boundary, or readiness question, which is `ai-governance`.
ai-governance
Use before rolling AI to a GTM team, or when the user says "AI policy", "shadow AI", "people are pasting customer data into ChatGPT", "AI governance", "are we allowed to use AI for this", "are we ready to use AI here", "the team is scared AI replaces them", "what data can go into the model", "works council", "AI compliance". Produces the usage policy, the data boundary, the disclosure rules, and the adoption plan that addresses fear rather than ignoring it. Writes workspace/agents/ai-policy.md. Sets policy for the team. Not for deciding whether one specific agent is worth building, which is `agent-scoping`.
artifact-eval
The output gate for this repo. Runs on any deliverable before it is handed over, and also when the user says "review this deliverable", "is this ready", "check this before I send it", "quality check". Verifies the artifact is complete, honest about its assumptions, and free of the specific failures each artifact type is prone to. Returns a receipt. Gates deliverables this repo produced. To score your own AI system output against a rubric, with a separate judge and a logged cycle, use `eval-loop`.
attribution-architecture
Use when the user says "attribution", "what did marketing source", "our numbers don't match sales", "multi-touch", "first touch last touch", "campaign influence", "prove marketing ROI", "which channel is working". Defines sourced, influenced, and accelerated with enforceable rules, sets the model and window before the campaign, and kills the double-counting. Writes workspace/reports/attribution-model.md.
context-portfolio
Use before building any agent that needs to know how a company operates, or when the user says "the agent doesn't know our positioning", "onboard the agent", "where does the messaging live", "our AI keeps getting the voice wrong", "context for the agent", "operating manual". Builds the ten-module written record every agent reads from, so positioning changes once and propagates everywhere. Writes workspace/agents/context/*.md.
credit-and-cost-control
Use before any run that spends money, or when the user says "credits are burning", "our enrichment bill is too high", "how much will this cost", "forecast credits", "compare providers", "set a spend limit", "we blew through our credits". Runs the pre-flight, forecasts the run, compares providers on cost per usable record, and installs ceilings that fail closed. Writes workspace/data/cost-model.csv.
data-hygiene
Use when the user says "clean this list", "we have duplicates", "dirty data", "data hygiene", "our CRM is a mess", "bounce rate is high", "stale records", "dedup", "normalize this". Runs the pre-paid cleanup pass, builds the ongoing decay and dedup loop, and produces a refresh queue driven by freshness rather than by calendar. Writes workspace/data/hygiene-report.md.
data-model-and-sync
Use before any list build, enrichment run, or CRM integration, or when the user says "what counts as an account", "our data model", "field mapping", "the sync is broken", "records are duplicating", "which system is the source of truth", "parent child accounts". Defines the object model, keys, field-level direction of truth, and sync rules. Writes workspace/data/object-model.md and a data dictionary.
enrichment-waterfall
Use when the user says "enrich this list", "find emails", "our match rate is low", "which provider should we use", "enrichment is expensive", "build a waterfall", "fill in these fields". Designs a cost-ordered multi-provider waterfall with a free gate in front, verification, graceful exit, and per-tier economics. Writes workspace/data/waterfall-spec.csv.
eval-calibration
Use after editing any eval spec, or when the user says "our scores keep going up", "is the rubric still right", "calibrate the spec", "rubric drift", "the eval passes everything now", "recalibrate". Regrades the frozen anchor set, measures drift, and distinguishes a generator that improved from a ruler that stretched. Writes workspace/evals/calibration/<type>-<date>.md. Tests the ruler, not the work. To judge an artifact use `eval-loop`, and to write or rewrite the rubric itself use `eval-spec-authoring`.
eval-loop
Use to run a quality check on any generated artifact, or when the user says "evaluate this", "score this draft", "is this good enough", "run the eval", "review this output", "check it against the rubric". Runs generate, judge, revise, re-judge, approve with the judge in a separate context so nothing grades its own work. Logs every cycle including failures. Writes to workspace/evals/log/. Scores your AI system output against a rubric you wrote. Not for writing that rubric (`eval-spec-authoring`), and not the pre-handoff gate on a deliverable this repo produced, which is `artifact-eval`.
eval-spec-authoring
Use before building anything whose output needs a quality bar, or when the user says "write a rubric", "eval spec", "how do we score this", "define what good looks like", "quality bar", "grading criteria", "our AI output is inconsistent". Encodes judgment as hard gates plus weighted dimensions with 1-to-5 anchors and a frozen calibration set. Writes workspace/evals/specs/<type>.md. Writes the rubric. Not for running it against an artifact (`eval-loop`) or retesting it against its anchors (`eval-calibration`).
gtm-systems-audit
Use at the start of any engagement, or when the user says "audit our systems", "where do we start", "what's broken in our GTM", "assess our stack", "we need to fix our operations", "new client onboarding". Scores all four layers 1 to 5, names the single constraint, and returns a sequenced fix plan with gates. Writes workspace/audit/systems-audit.md. Scores all four layers. Not a tool-by-tool inventory of what you pay for, which is `martech-stack-audit`.
icp-scoring-engine
Use when the user says "build a scoring model", "ICP score", "lead scoring", "our score doesn't work", "fit and intent", "which accounts should reps work", "MQL threshold", "prioritize this list". Builds fit and intent as separate axes, fits thresholds to the real distribution, and runs the discrimination test before shipping. Writes workspace/data/scoring-model.md.
lifecycle-and-routing
Use when the user says "lifecycle stages", "MQL definition", "lead routing", "round robin", "speed to lead", "leads are sitting untouched", "sales says the leads are bad", "SLA", "handoff", "who owns this lead". Defines stages as exit criteria, sets routing rules and SLAs with real measurement, and fixes the handoff that generates the recurring lead-quality argument. Writes workspace/reports/lifecycle.md.
martech-stack-audit
Use when the user says "audit our stack", "too many tools", "martech spend", "tool consolidation", "what should we cut", "renewal coming up", "our stack is too expensive", "do we need this tool", "CFO wants cuts". Inventories the stack, computes spend against revenue, finds redundancy and dead weight, and returns a KEEP / CONSOLIDATE / CUT / ADD verdict per tool with a defensible savings number. Writes workspace/stack/stack-audit.md.
ops-and-scale-os
The router for Ops & Scale OS. Use when the user says "ops and scale", "GTM engineering", "AI ops", "build the GTM system", "our data is a mess", or is not sure which skill they need. Reads the situation, names the constraint, and routes to exactly one skill. Also use to orient somebody new to the repo. Front door only. It routes and does not do the work: the audit itself is `gtm-systems-audit`, and scoping one agent is `agent-scoping`.
reporting-architecture
Use when the user says "our dashboards disagree", "build a scorecard", "board reporting", "what should we report", "metric definitions", "too many dashboards", "nobody trusts the numbers", "what's our source of truth for reporting". Builds the metrics dictionary with one definition and one owner per metric, then a layered reporting stack that stops the recurring definition argument. Writes workspace/reports/metrics-dictionary.csv.
signal-pipeline
Use when the user says "buying signals", "intent data", "trigger events", "job change alerts", "we get signals and nothing happens", "is this intent data worth it", "signal monitoring", "hiring triggers". Builds a signal taxonomy with decay windows, dedup, severity thresholds, and an explicit signal-to-play routing table. Includes the backfill discrimination test for vendor renewal. Writes workspace/data/signal-taxonomy.csv.
skill-authoring
Use when writing or fixing a Claude Code skill, or when the user says "write a skill", "my skill never triggers", "the wrong skill loads", "skill description", "turn this into a skill", "extract a skill", "the skill is too long". Covers anatomy, trigger design, negative boundaries, context economy, and the router pattern for large skill sets. Writes the SKILL.md.
tam-and-sourcing
Use when the user says "how big is our TAM", "source my market", "build the universe", "pull a list of companies", "market sizing", "where do we find these accounts", "we need more accounts in the funnel". Sizes the addressable universe, then pulls it in cost-controlled pages with dedup against what already exists. Writes workspace/data/universe.csv.
writing-quality
Runs automatically on any prose this repo produces that a human outside the team will read. Also use when the user says "does this sound like AI", "de-slop this", "clean up this copy", "check the writing", "make this sound human". Strips the machine tells, enforces the banned list, and returns the text with a receipt. Never skipped on external-facing prose.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.