agentgrasp MCP server
Commit Oct 1, 2026Manages long command outputs by saving them to files and providing tools for yes/no questions and relevant-file searching via TypeSafe Jev.
Also known as damejeras/agentgrasp
- Stars
- 0
- Tools
- —
- Downloads, 30 days
- —
- Latest version
- 0.1.0
- Last commit
- Oct 1, 2026
- License
- MIT
What is agentgrasp MCP?
agentgrasp manages long command outputs by saving them to files and providing tools for yes/no questions and relevant-file searching via TypeSafe Jev.
Summary generated from the project's README
About agentgrasp MCP server
Managing massive command outputs within a Claude Code agent's context window is what agentgrasp enables for developers. Instead of flooding the conversation with text, it intercepts long Bash command outputs, saves them to local files, and provides the agent with a concise summary and a file path. This approach keeps the context window clean while allowing the agent to interact with the data through specific tools.
Users interact with the data using the ask tool for yes/no questions and the search tool to identify relevant file segments and line ranges. Both tools rely on TypeSafe Jev, which requires a TYPESAFE_API_KEY set in the environment. The server runs locally and can be installed via a shell script or npm. Because it captures command output, Claude Code will prompt for approval of these grouped Bash commands unless running in bypassPermissions mode. Users must also ensure the filesystem sandbox allows writing to the agentgrasp state directory, typically located at ~/.local/state/agentgrasp.
What you can do with it
- 1Preventing context window overflowAn agent can run a long bash command and receive a summary instead of the full output to save space.
- 2Querying large files with yes/no questionsA developer can have an AI check for specific conditions in a log file using the ask tool.
- 3Locating specific information without reading everythingAn AI can find the exact lines in a file that matter to a query without pulling the entire file into the chat.
Written from the project's README. Every use case is checked against a quote from it.
Automated checks
Ran Nov 15, 2026Full report →Not verified yet. A server needs a score of 80, no serious finding, and enough data for us to check. How checks work · What to fix for more points
- Source repository linked and reachabledamejeras/agentgrasp is public and reachablegithub.com/damejeras/agentgrasp ↗Pass
- Package names this repository as its sourceNot checked
- Build provenance (Sigstore attestation)Not checked
- Open source license (OSI)MITspdx.org/licenses/MIT.html ↗Pass
- Commit or release in the last yearLast commit or release 2026-10-01github.com/damejeras/agentgrasp ↗Pass
- OpenSSF ScorecardNot checked
- Package name not close to a popular packageNot checked
- No critical advisory in the package or its dependenciesNot checked
- No hidden instructions to the model0 tool descriptions and the README read; no override phrases, credential paths or hidden charactersPass
- Destructive tools are labelledNot checked
- Remote endpoint answers the MCP handshakeNot checked
README
View on GitHubagentgrasp
agentgrasp keeps long command output out of a Claude Code agent's context. Claude Code still
runs every Bash command. When the output is long, agentgrasp saves it to a file and the agent
sees a short summary with the file path. The agent can then read the file, or ask yes/no
questions about it with the ask tool. The search tool finds the files and line ranges that
are relevant to a question, without returning source text.
ask and search use Jev, TypeSafe's yes/no model.
Install
curl --proto '=https' --tlsv1.2 -LsSf https://github.com/damejeras/agentgrasp/releases/latest/download/agentgrasp-installer.sh | sh
claude plugin marketplace add damejeras/agentgrasp
claude plugin install agentgrasp@agentgrasp
Instead of the installer, npm install -g agentgrasp installs the same binary, and
npx agentgrasp mcp runs the MCP server without a global install.
The plugin calls agentgrasp from PATH. Set TYPESAFE_API_KEY in the environment that
starts Claude Code; without it, ask and search return provider_unavailable.
Settings
agentgrasp runs each Bash command inside a { ... } group that sends its output to files.
Claude Code asks for approval of every command that holds such a group, also when your allow
rules allow the command, and no allow rule changes that. So in Claude Code's default and
acceptEdits permission modes, every Bash call asks for approval. In bypassPermissions mode
the calls run without a prompt. Your deny rules still block a command in every mode.
When the Claude Code sandbox is on, let Bash write the agentgrasp state directory. Without this, the output cannot go to its files, the shell prints the error, and the command does not run:
{
"sandbox": {
"filesystem": {
"allowWrite": ["~/.local/state/agentgrasp"]
}
}
}
Use $XDG_STATE_HOME/agentgrasp instead when XDG_STATE_HOME is set.
Costs and risks
askandsearchsend file content to TypeSafe. TypeSafe lists Jev at $42 per billion input tokens. Each call writes the tokens it used to its record under~/.local/state/agentgrasp/asks/orsearches/.askandsearchcan read any file inside the MCP roots, which are the project directory and the directories you add. This includes files that yourReaddeny rules block. When you allow these tools, you allow that.- The state directory grows until you delete files from it. agentgrasp never deletes them.
- A runaway command can fill the disk until Claude Code's timeout stops it.
Links
- MCP Registry name:
mcp-name: io.github.damejeras/agentgrasp
agentgrasp MCP questions
How do I set up the API key for agentgrasp?
You must set the TYPESAFE_API_KEY in the environment that starts Claude Code to enable the ask and search tools.
Does agentgrasp delete its state files automatically?
No, the state directory will continue to grow until you manually delete the files.
What happens if I use Claude Code with a sandbox enabled?
You must allow Bash to write to the agentgrasp state directory in your sandbox configuration, otherwise commands will fail.