Files and knowledge
Each agent in Agenta has two folders. Session files are temporary and belong to one session. Agent files persist and are shared across every session for that agent.
Session files support the work happening now
Suppose you upload this month's campaign spreadsheet and ask the marketing launch agent to explain why one audience performed better than another.
The spreadsheet belongs to that session. The agent can inspect it, create a chart, and revise the analysis with you. A new session does not start with that temporary working set.
Session files are useful for:
- Material you upload for one request.
- Drafts you are still shaping in the conversation.
- Downloads and intermediate results from a longer task.
- Finished files you want to take away, but do not need the agent to keep using.
Think of the session files as the desk for today's work.
Agent files carry useful context forward
Suppose the launch agent writes an email your team approves with one small edit. You want the next launch to benefit from that example.
Save the email in the agent's files, alongside the brand voice guide and approved product notes. Those files belong to the agent and are available across its sessions.
Agent files are useful for:
- Brand, product, and policy reference material.
- Examples of work your team approved.
- Finished reports worth consulting later.
- Notes and indexes that help the agent find its way around the collection.
Think of the agent files as the shared shelf the agent returns to from one conversation to the next.
| File scope | Best for | Launch agent example |
|---|---|---|
| Session | The task in front of you | This month's campaign spreadsheet |
| Agent | Work and knowledge worth reusing | The brand guide and approved launch email |
Patterns for using files
Wiki
You can build a wiki for your agent with Markdown files. Over time, it can collect research, plans, experiments, and instructions.
Write knowledge into files, then point at them from the instructions
To make the agent folder useful, write what the agent learns into files, then point to them from the instructions.
For example, after the marketing agent writes a few good launch emails, save them in examples/launch-emails/ in the agent folder. Then add one line to the instructions: "Before writing a launch email, read the files in examples/launch-emails/ and follow their structure."
A subfolder can carry its own AGENTS.md
You can put an AGENTS.md file inside a subfolder of the agent folder. When the agent works in that folder, the harness reads the file automatically along with the agent's main instructions.
A twitter-ads folder can carry its own rules for campaign names, spending limits, and report formats. An agent working elsewhere never loads those rules, which keeps the main instructions short.
A file adds nothing to the context until the agent opens it. This lets the agent's knowledge grow to hundreds of files without making every turn larger. Knowledge therefore belongs in files rather than in the instructions, which are sent on every turn.
Next
- Instructions covers how to point the agent at these files.
- Skills covers the other on-demand way to give an agent know-how.