hippo-feature

hippo-feature is a command for Claude Code from kitfunso/hippo-memory. It costs 17 tokens per session (1,217 once invoked), scanned A, original, MIT.

A command workflow for building one Hippo memory feature from research using a micro-evaluation TDD loop. TDD means writing a failing test first, then changing the code until it passes.

In plain words
What is it for?
Choosing a memory feature, writing a failing fixture, running quick micro-tests after each change, and progressing to larger LoCoMo tests before a pull request or release.
Why use it?
It sets required evaluation stages so memory changes are checked early and regressions are caught before larger or slower tests run.

Command for Claude Code

Install

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.

agentmods
npx agentmods add commands/kitfunso/hippo-memory/hippo-feature
Clone the repo
git clone --depth 1 https://github.com/kitfunso/hippo-memory

Made for: Claude Code.

Wrote 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.

agentmods badge for hippo-feature

README.md
[![agentmods](https://agentmods.dev/badge/commands/kitfunso/hippo-memory/hippo-feature.svg)](https://agentmods.dev/commands/kitfunso/hippo-memory/hippo-feature)
Your own site
<a href="https://agentmods.dev/commands/kitfunso/hippo-memory/hippo-feature"><img src="https://agentmods.dev/badge/commands/kitfunso/hippo-memory/hippo-feature.svg" alt="Measured on agentmods" height="20"></a>
Per session 17 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,217 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00017 $0.01217
Opus 5 $0.00009 $0.00609
Sonnet 5 $0.00003 $0.00243
Haiku 4.5 $0.00002 $0.00122

Measured 4d ago against content hash b27f875f6106, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

hippo-feature 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 4d 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.

.claude/commands/hippo-feature.md · 99 lines

How it starts

The opening of the file, as written. The whole thing — 99 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are entering hippo-feature mode to build a new memory mechanic from RESEARCH.md. Follow the eval pyramid strictly. Skipping tiers wastes hours of LoCoMo time.

The eval pyramid (non-negotiable)

Tier Harness Time When
1 benchmarks/micro/run.py ~30s Every code change
2 benchmarks/locomo/run.py --conversations 1 --sample 10 --score-mode evidence ~5-10 min Before opening a PR
3 LoCoMo full ~85 min evidence / ~6h judge Release gate only

If a feature shows no Tier 1 signal, do not proceed to Tier 2. If Tier 2 shows a regression, do not run Tier 3.

The loop

0. Pick the feature

If $ARGUMENTS names a feature (e.g. acc-evc, vmpfc-value, dlpfc-goals, vlpfc-gate, pineal-salience-v2), use it. Otherwise read the PFC priority table in RESEARCH.md (lines ~459-466) and propose the top three by effort × benchmark delta. Wait for user confirmation.

1. RED — Write the failing micro fixture FIRST

  • Drop a fixture at benchmarks/micro/fixtures/<feature>.json with shape:
    {
      "name": "<feature>-<aspect>",
      "mechanic": "<feature>",
      "description": "<one sentence: what behaviour this proves>",
      "remembers": [...],
      "queries": [{"q": "...", "must_contain_any": [...], "top_k": N}]
    }
    
  • The fixture must encode behaviour the current system cannot satisfy. If it passes on main today, the fixture is wrong — make it harder.
  • Run python benchmarks/micro/run.py --filter <feature> and confirm it fails.
  • Save the failing baseline: python benchmarks/micro/run.py --out benchmarks/micro/results/baseline-<feature>.json.

2. PLAN — Outside voice on non-trivial features

Per global CLAUDE.md outside-voice rule: if the feature touches schema, retrieval ranking, or storage (i.e. anything in the PFC priority table), run /plan-eng-review on the implementation plan before writing code. Skip only for one-line tweaks.

3. GREEN — Minimum implementation

Read the full file on GitHub · 99 lines

Changes

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.

  1. 4d ago First seen · 99 lines · 17 tokens per session scan A b27f875f6106

Subscribe to this mod's changes

hippo-feature is a command published in the GitHub repository kitfunso/hippo-memory (738 stars, last pushed 10d ago), licensed MIT. It adds 17 tokens to every session and 1,217 once invoked, about $0.0001 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.