Skip to content

Skills. Context. Usage.

Your agents' context.
Made visible.

ContextDaddy shows what the coding agents on your Mac can discover, invoke and consume. Review skill access, project context, local usage history and verified OpenTelemetry signals.

Public early-access beta · free to download · macOS 14+ · Apple silicon

Discovery is not invocation

Separate exposure from use.

ContextDaddy distinguishes what an agent could see from what it actually consumed. Installed cache evidence never masquerades as active runtime exposure, and context-size estimates stay visibly approximate.

Usage
Reads local agent-log history and provider allowance through installed CLIs.
Skills
Explains per-runtime discovery and invocation policy.
Projects
Shows discovered context files.
OpenTelemetry
Shows only verified signals from a compatible local collector.

Bounded by design

Reads metadata. Leaves prompts alone.

ContextDaddy reads bounded local metadata and configuration structure. It does not copy prompt bodies, tool arguments or credential values into its dashboards. Supported local skill and context-file changes require a preview and your explicit apply action.

For

people who run Codex, Claude, Cursor, Devin or Grok on a Mac and want to review skill access, context exposure and consumption, then explicitly preview supported local file changes.

Not for

live prompt inspection, automatic installation or activation, plugin-cache edits, cloud telemetry collection, or reconstructing usage that was never recorded.

Questions, answered.

Can I download ContextDaddy?
Yes. The beta is a free, signed and notarized Apple-silicon Mac build, distributed directly — no account, checkout, App Store or TestFlight listing. Updates are manual downloads for now.
Does it change my agent setup?
Discovery stays read-only. For supported local skill and context-file changes, ContextDaddy shows a preview and waits for your explicit apply action; it does not install skills automatically, edit managed plugin sources or activate skills in an agent runtime.
What does the usage view measure?
Local agent-log history via the bundled ccusage helper, provider allowance through installed provider CLIs when you ask, and OpenTelemetry signals from a compatible local collector. Missing sources are shown as unavailable, not zero activity.
Does anything leave my Mac?
Provider allowance checks use installed CLIs and OpenTelemetry reads a loopback collector. The app has no account, sync or telemetry upload of its own.
Which agents does it cover?
Discovery and policy models cover the major coding-agent runtimes, including Codex, Claude, Cursor, Devin and Grok. Non-auto skills show how to invoke them; policy is labelled explicit or derived.

Audit the access.
Trust the boundary.

Free during early access. Signed and notarized for Apple silicon on macOS 14 or later.

Illustrative artwork