← Back to Skills Marketplace
Terse
by
bowen31337
· GitHub ↗
· v2.0.0
· MIT-0
124
Downloads
0
Stars
0
Active Installs
4
Versions
Install in OpenClaw
/install terse
Description
Compress output by removing filler, pleasantries, articles, and hedging while preserving code, terms, and errors for concise technical responses.
Usage Guidance
This skill is basically a prompt-prefix/template for compressing sub-agent output and the included Python helper is harmless text-manipulation code. Before installing: 1) Confirm the platform will not apply 'terse' to owner-facing, planning, or security-sensitive outputs (the SKILL.md warns against that). 2) Ask the platform which 'always' source it honors — the SKILL.md embeds always: true while the registry shows always: false; ensure the skill will not be force-loaded. 3) Test on low-risk internal tasks (logs, CI steps, code implement/debug subtasks) to validate quality. 4) Review the linked repository yourself if you want extra assurance; there are no network calls or secret access in the included files. If you need help with specific checks (e.g., how your platform resolves SKILL.md metadata), share your platform details and I can guide next steps.
Capability Analysis
Type: OpenClaw Skill
Name: terse
Version: 2.0.0
The 'terse' skill is a utility designed to reduce token consumption by instructing sub-agents to use concise, 'caveman-style' language. The bundle consists of a helper Python script (scripts/caveman_prompt.py) for string formatting and markdown instructions (SKILL.md) that define safety boundaries and usage guidelines for the agent; no malicious code, data exfiltration, or unauthorized execution patterns were found.
Capability Assessment
Purpose & Capability
Name/description state a token-compression helper for agent responses and the package contains only a prompt-prefix generator and documentation that implement that feature. Required binaries/env/configs are none — proportionate to the described purpose.
Instruction Scope
SKILL.md focuses on generating compressed prompts and warns explicitly about 'hard exclusion' cases (planning, security, owner-facing output). That is appropriate, but the skill relies on the host/orchestrator to respect those exclusion rules; misuse (applying terse to critical outputs) would degrade quality. Also SKILL.md claims the skill 'auto-loads and applies to sub-agent responses' — ensure the platform enforces the stated exclusions.
Install Mechanism
No install spec and only a small, local Python helper script (caveman_prompt.py). No downloads from external URLs, no archive extraction, and no package manager pulls — low installation risk.
Credentials
No required environment variables, no credentials, and no config paths. The helper script is purely local string handling and does not access secrets or network resources.
Persistence & Privilege
Registry metadata shows always: false, but SKILL.md head includes metadata.openclaw: always: true (auto-classified). This is an inconsistency: if the platform honored the SKILL.md metadata and forced always-load, it would increase runtime presence and the blast radius of accidental misuse. Confirm which source the platform uses for the 'always' flag before enabling auto-application.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install terse - After installation, invoke the skill by name or use
/terse - Provide required inputs per the skill's parameter spec and get structured output
Version History
v2.0.0
- Added always-load metadata block to SKILL.md for auto-classification.
- No changes to logic or compression modes; documentation and core functionality remain the same.
v1.1.1
Remove personal identifiers from public docs. Generic 'owner-facing' terminology only.
v1.1.0
Harden exclusion rules: no terse for planning, architecture, writing, reviews, or user-facing comms. Explicit approved/excluded task lists.
v1.0.0
Initial release: compressed output mode for AI agents. 3 levels (lite/full/ultra), ~65-75% token savings. CLI helper included.
Metadata
Frequently Asked Questions
What is Terse?
Compress output by removing filler, pleasantries, articles, and hedging while preserving code, terms, and errors for concise technical responses. It is an AI Agent Skill for Claude Code / OpenClaw, with 124 downloads so far.
How do I install Terse?
Run "/install terse" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Terse free?
Yes, Terse is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Terse support?
Terse is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Terse?
It is built and maintained by bowen31337 (@bowen31337); the current version is v2.0.0.
More Skills