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/fockus/skill-memory-bankWrote 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/commands/fockus/skill-memory-bank/test)<a href="https://agentmods.dev/commands/fockus/skill-memory-bank/test"><img src="https://agentmods.dev/badge/commands/fockus/skill-memory-bank/test.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 | $0.00009 | $0.00540 |
| Opus 5 | $0.00005 | $0.00270 |
| Sonnet 5 | $0.00002 | $0.00108 |
| Haiku 4.5 | $0.00001 | $0.00054 |
Grade A, and why
test 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 3d 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 — 49 lines — stays where its author put it; the contents beside it link to each section on GitHub.
1. Delegate execution to mb-test-runner
Agent(
subagent_type="general-purpose",
model="sonnet",
description="mb-test-runner: run + parse project tests",
prompt="<contents of ~/.claude/skills/memory-bank/agents/mb-test-runner.md>
dir: ."
)
The agent detects stack via mb-metrics.sh, runs tests with per-stack parsing through scripts/mb-test-run.sh, and returns structured JSON: {stack, tests_pass, tests_total, tests_failed, failures[], coverage, duration_ms} plus a human summary. Use the JSON as the authoritative source for the rest of this flow.
If stack=unknown or the runner is missing, the agent reports tests_pass=null. Offer to create .memory-bank/metrics.sh (see references/templates.md).
If $ARGUMENTS provided a filter (test file, name, marker), pass it in the invocation context so the agent can narrow the run. Stage 3 of the runner does full-suite; filter support is follow-up work.
2. If tests_pass == true
Show the verdict + counts + duration from the agent's human summary. Done.
3. If tests_pass == false
For each failure (prioritize entries with touches_session == true):
- Read the failing test source and the code under test in full
- Check
.memory-bank/lessons.mdfor known flaky patterns or recurring anti-patterns - Classify: code bug vs. outdated test vs. environmental (flaky) issue
- Propose a concrete fix — show the diff
- Ask
y/Nbefore applying the fix (default = No)
The agent's failures[].likely_cause is a hint, not authority — always read the file yourself before proposing a fix.
4. If tests_pass == null (NOT-RUN)
Do not treat as pass. Report the reason (unknown stack, runner missing, zero tests collected) and offer the user a choice: install the runner, add .memory-bank/metrics.sh, or proceed without verification (flagged explicitly in the session log).
5. Memory Bank
If the run surfaced a recurring pattern worth remembering (flakiness, shared setup bug, environment issue), append to lessons.md using the template in references/templates.md.
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.
- 3d ago First seen · 49 lines · 9 tokens per session scan A 2a8429bc77e2
test is a command published in the GitHub repository fockus/skill-memory-bank (24 stars, last pushed yesterday), licensed MIT. It adds 9 tokens to every session and 540 once invoked, about $0.0000 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 2026-09-03.
Other commands, from other repositories
unity-ralph
Relentless verify-fix loop — refuses to stop until the project is clean. Runs unity-verifier repeatedly with configurable max iterations and stall detection.
unity-test
Writes missing tests and runs them via MCP. Identifies untested code, creates EditMode/PlayMode tests, executes via runtests, reports results.
build-apex
Wraps agents/apex-builder/AGENT.md. Produces Apex class(es) + test class that conform to templates/apex/ (TriggerHandler, BaseService, BaseSelector, ApplicationLogger, SecurityUtils, HttpClient).
build-flow
Wraps agents/flow-builder/AGENT.md. Returns the Flow type decision, element plan, subflow plan, fault path, bulkification notes, and test matrix.
gen-tests
Wraps agents/test-class-generator/AGENT.md. Takes a target Apex class, produces a test class targeting ≥ 85% coverage using the canonical factories in templates/apex/tests/.
refactor-apex
Wraps agents/apex-refactorer/AGENT.md. Takes a target class, returns a refactored version + test class using templates/apex/.