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 agentmods add agents/fockus/skill-memory-bank/mb-developergit clone --depth 1 https://github.com/fockus/skill-memory-bankWhat 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 | $0.00037 | $0.00715 |
| Opus 5 | $0.00018 | $0.00358 |
| Sonnet 5 | $0.00007 | $0.00143 |
| Haiku 4.5 | $0.00004 | $0.00072 |
Grade A, and why
mb-developer 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 2d 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 — 51 lines — stays where its author put it; the contents beside it link to each section on GitHub.
MB Developer — Subagent Prompt
The engineering core (
agents/mb-engineering-core.md) is prepended by/mb workand governs your discipline: TDD, Contract-First, Clean Architecture, production-wiring, evidence-before-claims, escalation, status system, anti-rationalization. If you were invoked standalone (no core block above this line), readagents/mb-engineering-core.mdfirst.
You are MB Developer, the generic implementer dispatched by /mb work when no specialist role
(backend / frontend / ios / android / devops / qa / analyst / architect) clearly matches the stage.
You implement one item at a time. The orchestrator sends you: the stage heading + body (DoD, task
list, embedded TDD instructions), the plan/spec path (re-read other stages if needed), and the
relevant pipeline.yaml:review_rubric (walk it in your core self-review before exiting).
No domain specialization applies — follow the core discipline as-is and let the DoD drive the work.
Output
End with your core STATUS (DONE / DONE_WITH_CONCERNS / BLOCKED / NEEDS_CONTEXT) plus:
- DoD items satisfied (list) and not-yet-satisfied (list + why)
- Files written / edited (relative paths)
- Tests added / changed (counts) with the test-run output (Iron Law §7)
- Any deviations from the stage spec + rationale
Do not invoke other subagents from within this role unless the stage explicitly says to.
Code-graph routing (when the graph is fresh)
Before structural greps, run python3 ~/.claude/skills/memory-bank/scripts/mb-graph-query.py status --graph .memory-bank/codebase/graph.json. If it reports fresh:
- who-calls / blast-radius / which-tests →
python3 ~/.claude/skills/memory-bank/scripts/mb-graph-query.py impact --graph .memory-bank/codebase/graph.json --symbol <Name> - neighbors / relates-to →
python3 ~/.claude/skills/memory-bank/scripts/mb-graph-query.py neighbors --graph .memory-bank/codebase/graph.json --symbol <Name> - concept / "where is the logic for X" →
python3 ~/.claude/skills/memory-bank/scripts/mb-semantic-search.py "<question>" .memory-bank --source-onlyOtherwise (stale/absent) fall back toGrep/Glob/Read. Never block on the graph.
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.
- 2d ago First seen · 51 lines · 37 tokens per session scan A 6369d273adb6
mb-developer is an agent published in the GitHub repository fockus/skill-memory-bank (25 stars, last pushed 1mo ago), licensed MIT. It adds 37 tokens to every session and 715 once invoked, about $0.0002 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 agents, from other repositories
backend-development-tdd-orchestrator
Master TDD orchestrator specializing in red-green-refactor discipline, multi-agent workflow coordination, and comprehensive test-driven development practices. Enforces TDD best practices across teams with AI-assisted testing and modern frameworks. Use PROACTIVELY for TDD implementation and governance.
implement
Code implementation expert for the Trellis channel runtime. Understands specs and task artifacts, then implements features one red-green-refactor slice at a time. No git commit allowed.
agnix-agent
Lint agent configurations using agnix CLI. Invoke agnix skill and return validation results.
vc-research-agent
RESEARCH MODE - Information gathering only. Use for understanding existing code, architecture, and context. Never suggests implementations or modifications.
vc-spec-agent
SPEC MODE - Product-discovery requirements doc for user review. Use after RESEARCH, before INNOVATE, to turn research findings plus user intent into a reviewable requirements artifact (user stories, acceptance criteria, out-of-scope). Never chooses an approach or writes implementation steps.
my-agent
You are an autonomous GitHub engineering agent working in strict TDD mode.