clause-logic/exoclaw agent
Run the exoclaw AI agent in response to GitHub events
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Stale
- License
- None
Inputs
| name | description | required | default |
|---|---|---|---|
| model | LLM model to use. Defaults to github/gpt-4.1-mini which uses your GITHUB_TOKEN — no extra secrets needed. To use Anthropic or OpenAI, set e.g. claude-sonnet-4-6 or gpt-4o and provide the relevant API key secret. | no | github/gpt-4.1-mini |
| trigger | Word that must appear in comments to trigger the bot. Empty string = respond to all. | no | @exoclawbot |
| respond_to_issues_opened | Respond when an issue is opened | no | true |
| respond_to_prs_opened | Respond when a PR is opened | no | false |
| python_version | Python version to use | no | 3.11 |
| tools | Comma-separated list of tools to enable. Workspace tools (read/write/exec) are always on. GitHub: github_review, github_label, github_pr_diff, github_issue, github_reaction, github_file, github_checks, github_search. Web: web_fetch (fetch a URL as Markdown), web_search (needs a web-search capable model/provider). Default: none enabled. | no | "" |
| token_budget | Per-turn token budget. The agent gets progressive warnings and a graceful cutoff as it approaches this ceiling, keeping a single run's cost bounded. | no | 1500000 |
| package_version | exoclaw-github version to install (default latest) | no | "" |
Outputs
no outputs