yzj623/pplab/apt-simulation-proposal-protocol
Claude Code and agent skill by yzj623/pplab
- community
Checking that the SKILL.md file is still in the repository.
Run APT-style sequential simulator proposal rounds with auditable prior, proposal, observation, and update records for likelihood-free inference recovery.
Install
npx vskill@latest install yzj623/pplab/apt-simulation-proposal-protocolSKILL.md
The same SKILL.md was published earlier as vectorspacelab/arex-skill/apt-simulation-proposal-protocol.
APT Simulation Proposal Protocol
Use this skill when you need to set up or audit a bounded Automatic Posterior Transformation recovery experiment that draws simulator parameters from a prior or adaptive proposal and records which proposal generated each simulation. Do not use it as a posterior-density trainer by itself; pair it with an APT transformation or atomic-loss skill for training.
Inputs
- A prior distribution with a name, parameters, sampler, log density, and optional support check.
- A simulator callable or deterministic simulator specification.
- Observed data
x_o. - A current posterior estimate at
x_ofor proposal updates. - Round count, simulations per round, and deterministic seed.
Outputs
- Round records with
round_index,theta,x,proposal_name,proposal_parameters, andseed. - Proposal update records linking
proposal_rtoproposal_{r+1}. - Source metadata showing the simulator and proposal parameters used for each item.
Workflow
- Start with
proposal_1 = prior. - For each round, draw parameters from the current proposal and run the simulator.
- Preserve the proposal parameters inside every simulation record.
- Train or update the posterior estimator using APT loss outside this skill.
- Construct the next proposal from the posterior estimate at
x_o. - If the prior has bounded support, reject or clip out-of-support proposal samples and record the policy.
Deterministic Helper
The script scripts/sequential_protocol.py provides a scalar Gaussian helper for reduced recovery:
It simulates a first round from a broad prior and a second round from a posterior-derived proposal. The helper is intentionally small; its purpose is to validate proposal provenance and sequential update mechanics, not to reproduce paper-scale neural training.
Validation
Run:
The tests verify first-round prior use, second-round proposal use, deterministic records, and valid support handling.
Limitations
- The helper supports scalar Gaussian examples only.
- It does not implement MDN, MAF, CNN, or RNN density estimators.
- It does not claim full paper recovery unless combined with executable APT loss and evaluation evidence.
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
Popularity
- 7-day trend score
- 5
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/yzj623/pplab/apt-simulation-proposal-protocol)