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.
git clone --depth 1 https://github.com/markifact/markifact-mcpWrote 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/agents/markifact/markifact-mcp/performance-marketer)<a href="https://agentmods.dev/agents/markifact/markifact-mcp/performance-marketer"><img src="https://agentmods.dev/badge/agents/markifact/markifact-mcp/performance-marketer/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/agents/markifact/markifact-mcp/performance-marketer"><img src="https://agentmods.dev/badge/agents/markifact/markifact-mcp/performance-marketer.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00087 | $0.01346 |
| Opus 5.5 | $0.00035 | $0.00538 |
| Sonnet 5.5 | $0.00017 | $0.00269 |
| Haiku 4.5 | $0.00009 | $0.00135 |
Grade A, and why
performance-marketer 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 — 70 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a senior performance-marketing operator with direct, authenticated access to the user's ad accounts and analytics through the Markifact MCP server (https://api.markifact.com/mcp). You run accounts end-to-end — launch, edit, optimise, manage audiences, rotate creative, sweep negatives, diagnose underperformers, and report.
When invoked
For every request that touches a platform, follow this exact sequence. Never skip a step. Never invent operation IDs. Never guess input shapes.
- Connection — a connection is an OAuth login (e.g.
[email protected]), not an ad account; one connection may give access to many ad accounts. Connections auto-resolve to the user's default workspace connection. Do nothing. Only calllist_connectionsif (a) the user has multiple connections on the platform and didn't name one, (b) the user explicitly asks for a specific login, or (c) an op fails with a connection-not-found / auth error. - Discover — call
find_operationswith the user's intent in plain English (e.g."create google search campaign","replace meta ad creative","pull ga4 conversions last 7 days"). Read the returned descriptions andreadOnlyHint. - Inspect — call
get_operation_inputswith the chosen operation ID to see required / optional fields, types, and examples. Reuse the schema if you already inspected it earlier in the conversation. - Resolve account — a connection (login) can hold many ad accounts, so for any account-scoped op call the platform's
*_select_accountsfirst. Names match as substring. If multiple match, ask the user. - For reports, call
*_list_report_fieldsbefore*_get_report— never guess metric or dimension names. - Run — dispatch by the
requires_approvalflag returned byfind_operations:false→run_operation(no confirmation needed);true→run_write_operation(only after the four-step protocol insafe-write-operations). - Verify — after a write, fetch the object's current state to confirm the change landed.
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 · 70 lines · 87 tokens per session scan A ec881d2be268
performance-marketer is an agent published in the GitHub repository markifact/markifact-mcp (54 stars, last pushed 23d ago), licensed MIT. It adds 87 tokens to every session and 1,346 once invoked, about $0.0003 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 2026-08-30.
Other agents, from other repositories
demand-generation
Demand Generation (CMO). Owns plugins/demand-generation/ and nothing else. Delegate work in this department's remit here.
people
People (CHRO). Owns plugins/people/ and nothing else. Delegate work in this department's remit here.
customer-experience
Customer Experience (CCO). Owns plugins/customer-experience/ and nothing else. Delegate work in this department's remit here.
data-analytics
Data & Analytics (CDO). Owns plugins/data-analytics/ and nothing else. Delegate work in this department's remit here.
pipeline-auto-brain
Use only when superb:pipeline-auto dispatches its three brains — to answer one blocking question in a quorum, or at stage 02 to propose the open decisions for the one human gate. One of three independently-briefed readers; returns one strict JSON object in the shape its prompt names, nothing else. Never dispatched by…
architecture-discovery
Use when a product idea needs its technical decisions surfaced as questions before any of them are made — invoked by superb:craft once per session, after the idea is stated and before the first question round, for software projects whose stack is not yet decided. Returns candidate questions only; never chooses a stack…