How do I install the Eld Record Memory Collector skill?

Run npx skillmds@latest add majiayu000/eld-record-memory-collector-2 in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Eld Record Memory Collector safe to use?

This skill has not completed SkillMD's automated safety review yet. Independent scanners report: SkillSpector: PASS, Skill Scanner: PASS. Capability flags: docs only. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Eld Record Memory Collector?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Eld Record Memory Collector free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Eld Record Memory Collector?

majiayu000 (@majiayu000) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Eld Record Memory Collector by majiayu000

Name: Eld Record Memory Collector
Rating: 4.2 (5 reviews)
Author: majiayu000

Eld Record Memory Collector

Collect and verify project information for pce-memory with multi-model validation using Codex CLI. Use when gathering project context, onboarding to a new codebase, or building a knowledge base. Triggers on "collect project info", "build knowledge base", "index this project", or "remember this codebase".

majiayu000 Updated Sep 22, 2026 567 repo stars

File contents

PCE Memory Collector

Collect project information with multi-model verification to ensure accuracy.

Workflow

Loop 1: Collect   →  observe(ttl=7, unverified)
Loop 2: Verify    →  Codex cross-check → observe+extract(verified)
Loop 3: Validate  →  hash/diff check → feedback

Loop 1: Initial Collection

Scan project structure:
```
scripts/scan_project.py <project_path>
```

Read key files and record with observe:

observe(
  source_type="file",
  content="<extracted info>",
  source_id="<file_path>",
  ttl_days=7,
  boundary_class="public|internal",
  tags=["unverified", "<category>"]
)

Categories: project-info, architecture, dependencies, api, config

Loop 2: Multi-Model Verification

Cross-validate with Codex CLI:

mcp__codex-cli__codex(
  prompt="Verify this claim against the file content:
    Claim: <claim_text>
    File: <file_path>
    Content: <file_content>

    Respond: MATCH | MISMATCH | PARTIAL
    Reason: <brief explanation>"
)

Decision Matrix:

Codex Result	Action
MATCH	`observe(..., extract={mode: "single_claim_v0"})` → permanent
PARTIAL	Refine claim, re-verify
MISMATCH	Discard (let ttl expire)

Loop 3: Hash-Based Validation

Run validation script:

scripts/validate_claims.py <project_path>

The script:

Activates all project claims
Computes current file hashes
Compares with stored provenance
Outputs: VALID | OUTDATED | MISSING

Send feedback:

feedback(claim_id, signal="helpful|outdated", score)

Quick Start

1. scan_project.py /path/to/project     # Discover files
2. [Manual] Read files, observe         # Loop 1
3. [Codex] Cross-validate               # Loop 2
4. validate_claims.py /path/to/project  # Loop 3

Scripts

scripts/scan_project.py - Scan project and output key files
scripts/validate_claims.py - Hash-based claim validation

References

references/pce_memory_api.md - pce-memory tool reference

---name: eld-record-memory-collector-2description: Collect and verify project information for pce-memory with multi-model validation using Codex CLI. Use when gathering project context, onboarding to a new codebase, or building a knowledge base. Triggers on "collect project info", "build knowledge base", "index this project", or "remember this codebase".---# PCE Memory CollectorCollect project information with multi-model verification to ensure accuracy.## Workflow```Loop 1: Collect → observe(ttl=7, unverified)Loop 2: Verify → Codex cross-check → observe+extract(verified)Loop 3: Validate → hash/diff check → feedback```## Loop 1: Initial Collection1. Scan project structure: ```bash scripts/scan_project.py <project_path> ```2. Read key files and record with observe: ``` observe( source_type="file", content="<extracted info>", source_id="<file_path>", ttl_days=7, boundary_class="public|internal", tags=["unverified", "<category>"] ) ```3. Categories: `project-info`, `architecture`, `dependencies`, `api`, `config`## Loop 2: Multi-Model VerificationCross-validate with Codex CLI:```mcp__codex-cli__codex( prompt="Verify this claim against the file content: Claim: <claim_text> File: <file_path> Content: <file_content> Respond: MATCH | MISMATCH | PARTIAL Reason: <brief explanation>")```**Decision Matrix:**| Codex Result | Action ||--------------|--------|| MATCH | `observe(..., extract={mode: "single_claim_v0"})` → permanent || PARTIAL | Refine claim, re-verify || MISMATCH | Discard (let ttl expire) |## Loop 3: Hash-Based ValidationRun validation script:```bashscripts/validate_claims.py <project_path>```The script:1. Activates all project claims2. Computes current file hashes3. Compares with stored provenance4. Outputs: `VALID | OUTDATED | MISSING`Send feedback:```feedback(claim_id, signal="helpful|outdated", score)```## Quick Start```1. scan_project.py /path/to/project # Discover files2. [Manual] Read files, observe # Loop 13. [Codex] Cross-validate # Loop 24. validate_claims.py /path/to/project # Loop 3```## Scripts- `scripts/scan_project.py` - Scan project and output key files- `scripts/validate_claims.py` - Hash-based claim validation## References- `references/pce_memory_api.md` - pce-memory tool reference

Frequently asked questions