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 featbit/featbit-skills --skill featbit-experimentationgit clone --depth 1 https://github.com/featbit/featbit-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/featbit/featbit-skills/featbit-experimentation)<a href="https://agentmods.dev/skills/featbit/featbit-skills/featbit-experimentation"><img src="https://agentmods.dev/badge/skills/featbit/featbit-skills/featbit-experimentation.svg" alt="Measured on agentmods" height="20"></a>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 | $0.00066 | $0.07138 |
| Opus 5 | $0.00033 | $0.03569 |
| Sonnet 5 | $0.00013 | $0.01428 |
| Haiku 4.5 | $0.00007 | $0.00714 |
Grade A, and why
featbit-experimentation 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 7d 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 — 562 lines — stays where its author put it; the contents beside it link to each section on GitHub.
FeatBit Experimentation
This skill is the single entry point for the FeatBit release-decision loop:
intent -> hypothesis -> implementation -> exposure -> measurement -> interpretation -> decision -> learning -> next intent
It replaces the old multi-skill layout (featbit-release-decision, intent-shaping, hypothesis-design, reversible-exposure-control, measurement-design, experiment-workspace, evidence-analysis, learning-capture, and project-sync). Do not call those split skills from this workflow. Treat their former behavior as internal stages of this skill.
Core Principles
- Do not let an available tool define the user's problem before the decision type is clear.
- Do not implement before the business intent and hypothesis are explicit.
- Do not expose a measurable change before reversibility and rollout rules are defined.
- Do not measure with multiple primary metrics. One metric decides; guardrails protect.
- Do not interpret experiment data before sufficiency, SRM, instrumentation, and window checks pass.
- Do not close a cycle without recording a reusable learning and a next hypothesis.
- Persist every stage transition through the configured FeatBit experimentation MCP server.
MCP Contract
The FeatBit API database is the canonical source for experiment state. Read it on entry and write it before moving stages.
Required MCP tools:
| Tool | Purpose |
|---|---|
featbit_experiment_get_experiment |
Read experiment state, runs, messages, setup mode, and pasted input data |
featbit_experiment_update_experiment |
Write goal, intent, hypothesis, constraints, lastAction, lastLearning |
featbit_experiment_set_stage |
Set lifecycle stage |
featbit_experiment_update_metrics |
Write primary metric and guardrails |
featbit_experiment_create_run |
Create an experiment run |
featbit_experiment_update_run |
Update run setup, status, decision, and learning fields |
featbit_experiment_update_run_traffic |
Configure a run's experiment traffic assignment: analysis method, control/treatment roles, layer id/key, bucket slice, assignment unit, audience filters, allocation plan, and per-variation analysis sampling |
featbit_experiment_analyze_run |
Run server-side analysis and persist inputData / analysisResult |
featbit_experiment_list_layers |
List registered release-decision layers for the experiment environment |
featbit_experiment_create_layer |
Create a registered layer after explicit user approval |
featbit_experiment_update_layer |
Update a registered layer after explicit user approval |
featbit_experiment_archive_layer |
Archive a registered layer after explicit user approval |
featbit_experiment_get_feature_flag |
Read the real FeatBit flag, revision, variations, and targeting for the experiment environment |
featbit_experiment_create_feature_flag |
Create a FeatBit-managed feature flag after the exposure contract is complete and the user explicitly approves |
featbit_experiment_update_feature_flag_targeting |
Update flag targeting/rollout directly, or create a change request when useChangeRequest or reviewers are provided, after explicit user approval |
featbit_experiment_toggle_feature_flag |
Enable or disable the FeatBit flag after targeting is configured, or during pause/rollback execution, after explicit user approval |
What ships with it
21 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.
- references/evidence-decision-framing-guide.md 6.6 KB
- references/evidence-tool-featbit-abtesting.md 3.9 KB
- references/exposure-mcp-lifecycle.md 4.7 KB
- references/exposure-multi-experiment-traffic.md 7.5 KB
- references/exposure-pm-dev-handoff.md 5.1 KB
- references/exposure-rollout-patterns.md 3.5 KB
- references/exposure-tool-featbit-webui.md 5.3 KB
- references/hypothesis-template.md 3.0 KB
- references/intent-goal-extraction-patterns.md 2.7 KB
- references/learning-iteration-synthesis-template.md 5.5 KB
- references/mcp-tool-usage.md 9.0 KB
- references/measurement-event-schema-design.md 6.0 KB
- references/measurement-tool-featbit-sdk.md 3.4 KB
- references/ui-stage-prompts.md 9.4 KB
- references/workspace-analysis-bandit.md 10.0 KB
- references/workspace-analysis-bayesian-decision-guide.md 9.2 KB
- references/workspace-analysis-bayesian-usage-patterns.md 7.0 KB
- references/workspace-analysis-bayesian.md 10 KB
- references/workspace-analysis-holdout.md 6.6 KB
- references/workspace-data-source-guide.md 7.0 KB
- references/workspace-experiment-folder-spec.md 8.7 KB
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.
- 7d ago First seen · 562 lines · 66 tokens per session scan A 0c95bfa8fdac
featbit-experimentation is a skill published in the GitHub repository featbit/featbit-skills (11 stars, last pushed 3d ago), licensed MIT. It adds 66 tokens to every session and 7,138 once invoked, about $0.0003 per session on Opus 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 2026-08-30.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
next-partial-prefetching-adoption
Turn on Partial Prefetching in a Next.js app and work through the insights it surfaces. Use when the user wants to enable or adopt Partial Prefetching, flip the partialPrefetching flag, opt routes in with export const prefetch = 'partial', audit Link prefetch={true} behavior, preserve existing prefetched UI with…
chronicle
Analyze Copilot session history for standup reports, usage tips, session search, and session reindexing. Use when the user asks for a standup, daily summary, usage tips, workflow recommendations, wants to search or find past sessions by keyword/file/PR, wants to reindex their session store, or asks about deleting…
babysit-pr
Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…