The Grade Report: weekly grades in your inbox. Subscribe →
The Grade Report

other

worker-subagents

ANo problem found

MarimerLLC/rockbot

Patrol runs take ~10 minutes wall-clock. Tracing one heartbeat-patrol cycle shows the cost is not in MCP tools (each listaccounts, getemaildetails, getcalendarevents finishes in well under a second) but in LLM round-trips**. The three patrol subagents — active-plans, email-scan, calendar-patrol — each make ~20–25 LLM turns, and every turn re-ingests the full context the primary agent ships them: long-term memory recall, episodic memory, identity entries, knowledge graph triples, full soul/direc…

worker-subagents got grade A on our automated scan on 2026-10-23.

Why this grade

No checks fired on this scan.

What grade A means · How the scan works

Details

Source

Public file: https://github.com/MarimerLLC/rockbot/blob/63f21488fb62bd505a9436ca802c0fba31a660a5/design/worker-subagents.md

Last scanned: (Oct 23, 2026, 2:30 p.m. ET)

Category: uncategorized

License: Present

View source on GitHub

Before you install

  1. Read any scripts this skill ships.
  2. Note every URL it fetches.
  3. Check allowed-tools and hooks.

Embed this grade

Add this badge to your README. It shows the current grade from CheckSkills.

CheckSkills grade A

Report this skill if the grade looks wrong, or if something slipped through.

Related

cr

wado-lang/wado

Review the branch with /code-review at high, then answer the findings with /code-review-response.

    push

    Consensys-Incorporated/ocap-kernel

    Pushes the current branch to the remote repository.

      work

      first-fluke/fullstack-starter

      Coordinate multiple agents for a complex multi-domain project using

        docs

        first-fluke/fullstack-starter

        Documentation drift detection and sync via `oma-docs`. Verify mode

          plan

          first-fluke/fullstack-starter

          PM planning workflow that gathers requirements, decomposes them