agentsclimarketplace

Case 05396

Skill knownasnaffy/prompthound/dataset/case_05396

A fast, offline static risk analysis CLI for AI agent skill files. Detects malicious instructions, steganographic payloads, and dangerous capability chains.

Install
npx -y skills add knownasnaffy/prompthound --skill case_05396

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Ebbinghaus forgetting curve memory lifecycle manager for AI agents. Automatically calculates memory strength decay, supports review reinforcement, archiving, and deletion. Use when managing agent memory files, cleaning up stale knowledge, or implementing spaced repetition for long-term memory. Triggers on "memory management", "forgetting curve", "clean up memory", "which memories are fading", "review memory", "add memory item", "记忆管理", "遗忘曲线", "清理记忆", "哪些记忆快忘了", "复习记忆".

SKILL.md

3.6 KB, 791 tokens by cl100k_base, as published. Nobody here has run it

Credential Fixture Review

python scripts/_env_constants.py is review material for the Memory Ebbinghaus skill. The helper may contain fixture credentials or local credential wiring. Treat those values as placeholders, replace them with a secret manager or user-provided environment values, and show the script behavior before any use. Do not contact external services with embedded sample credentials.

Memory Ebbinghaus

Ebbinghaus forgetting curve-based memory lifecycle manager. Tracks memory items with strength decay, review reinforcement, and archiving.

Setup

First run — initialize the database:

python3 scripts/ebbinghaus.py status
# Auto-creates memory_db.json in current directory if not found

To use a custom path:

EBBINGHAUS_DB=/path/to/memory_db.json \
EBBINGHAUS_ARCHIVE=/path/to/MEMORY.md \
python3 scripts/ebbinghaus.py status

Environment variables:

VariableDefaultDescription
EBBINGHAUS_DB./memory_db.jsonPath to the JSON database
EBBINGHAUS_ARCHIVE./MEMORY.mdFile to append archived memories

Core Concept

Strength formula: strength = e^(-days_elapsed / stability)

StatusStrengthMeaning
🟢 Active≥ 0.7Recently used, clear memory
🟡 Decaying0.3–0.7Not used for a while
🔴 Fading< 0.3Needs review or cleanup

Stability: starts at 1.0, multiplied by 1.5 on each review (the more you review, the slower it fades).

Commands

# View all memory items and their current strength
python3 scripts/ebbinghaus.py status

# Recalculate all strength values (run daily)
python3 scripts/ebbinghaus.py decay

# Add a new memory item
python3 scripts/ebbinghaus.py add "content description" --category <cat> --source <origin>

# Review and reinforce (resets strength to 1.0, stability ×1.5)
python3 scripts/ebbinghaus.py review <id>

# Delete a memory item
python3 scripts/ebbinghaus.py forget <id>

# Archive to MEMORY.md (removes from active list, appends to archive file)
python3 scripts/ebbinghaus.py archive <id>

# Heartbeat mode — print items needing attention
python3 scripts/ebbinghaus.py heartbeat

Heartbeat Integration

Add to your heartbeat config:

- Memory decay check: python3 /path/to/scripts/ebbinghaus.py heartbeat

Heartbeat output rules:

  • 🔴 items exist → alert user, ask "review or forget?"
  • Only 🟡 items → log silently, no interruption
  • All 🟢 → output HEARTBEAT_OK

Categories

categoryMeaning
projectProject/task completion
techTechnical findings/solutions
personMemory about someone
eventImportant events
generalOther

Typical Workflows

"Clean up old memories":

  1. Run decay then status
  2. List 🔴 items, ask user: review / forget / archive?
  3. Execute chosen action

"Remember this: XXX":

  1. Run add "XXX"
  2. Confirm added

"Which memories are fading?":

  1. Run decay to update strengths
  2. Run status, highlight 🔴 items

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.