How to use the EIIRP Post-Work Organizer agent skill

What the EIIRP skill makes the agent do

After a significant work session, EIIRP makes the agent run a seven-phase audit. The phases are: inventory every output, walk the taxonomy to decide where each piece lands, check that the brain's schema still fits its actual shape, file enriched pages with cross-links, audit the skill set for duplication and gaps, verify that links resolve, and report back.

The skill also includes an always-on rule. When the agent is about to deliver a long analysis (500 words or more) of a document you shared, it saves that analysis as a brain page first and includes the link in the same reply, so the work is not lost in chat.

• Inventory: list every output produced in the session.
• Taxonomy: decide where each piece of knowledge belongs in the brain.
• Schema check: confirm the brain's structure still fits the content.
• Filing: write enriched pages with cross-links.
• Skill audit: look for reusable patterns and overlapping or duplicate skills.
• Verification: confirm links resolve.
• Report: summarise what was filed and where.

Install the EIIRP skill

Download the ZIP file from this page. The skill is a single file, SKILL.md, inside the plugin's skills/eiirp folder. There are two ways to install it.

For Claude Code, unzip the download into your personal skills folder, ~/.claude/skills/, or into a project's .claude/skills/ folder if you want it limited to one project.

For the Claude apps, upload the ZIP under Settings → Capabilities → Skills.

• Personal install: unzip into ~/.claude/skills/.
• Project install: unzip into the project's .claude/skills/ folder.
• Claude apps: upload the ZIP under Settings → Capabilities → Skills.

Check the requirements before you trigger it

EIIRP writes to a gbrain knowledge base. It needs the gbrain CLI, because the skill runs 'gbrain schema show --json' to read the active schema pack. It also needs these brain tools: search, query, get_page, put_page, add_link and add_timeline_entry.

If you do not have a gbrain knowledge base, this skill is not useful yet. It writes into brain directories and assumes that structure already exists.

• gbrain knowledge base with the gbrain CLI installed.
• Brain tools: search, query, get_page, put_page, add_link, add_timeline_entry.
• The skill runs no scripts and makes no network calls, according to its listing.

Trigger the skill with natural requests

You do not need a special command. The skill responds to phrases such as the ones below. Use them after a work session, when you want output filed, or when you want skills checked for overlap.

The skill can also fire on its own. The auto-filing rule runs when you share a document and the agent is about to deliver a substantial analysis of it.

• "Everything in its right place" or "EIIRP"
• "Store this research" or "put this in the brain"
• "File this properly" or "where does this research go"
• "Make this permanent" or "archive this research thread"
• "DRY this up" or "check everything is in the right place"
• "Analyze this document" or "deep analysis"

Example: filing a document analysis

Suppose you share a long PDF and ask the agent to analyse its findings. If the analysis runs to 500 words or more of structured findings, the always-on rule applies. The agent should file the analysis as a brain page first, then reply with the analysis and a link to that page in the same message.

You can also ask directly after a session, for example: "File all of this work properly." The agent then runs the seven-phase audit and reports which pages it created or updated.

• Tell the agent to file a finished research thread, and it runs the full audit.
• Expect a report listing the pages created or updated and any links it checked.

Safety notes: writes, scripts and sensitive documents

EIIRP writes pages to your brain. The auto-filing rule can write pages without an explicit request, and it fires on any user-shared document of sufficient length. If a document is confidential, do not share it with the agent when you do not want its analysis stored in a persistent store, or turn auto-filing off before you share it.

The skill runs no scripts and uses no network access, according to its listing. Its writes go to the brain through the tools listed above.

• The skill is marked as mutating and writes pages.
• Auto-filing can save analyses you did not ask to keep.
• Quick questions and simple lookups should be skipped by the rule, but the agent may misjudge them and file them anyway.

Tips for better results

Run EIIRP at the end of a session, not after every message, so the audit sees the full set of outputs. Give the agent a clear statement of what the session produced.

If you maintain several agent skills, run the audit after building a new one. It checks for overlap with existing skills and for gaps in coverage.

• Run the audit once per completed thread, not per message.
• Name the output you care about so the inventory step does not miss it.
• Use the skill-overlap check when you add new skills to a shared library.