juanfranpaezz

Selfcert Guard — Development agent for Claude Code

Development community

Use when code claims its own correctness and needs an intent-only, body-blind executable postcondition check.

How to install Selfcert Guard

Installs to ~/.claude/agents/juanfranpaezz-harness-selfcert-guard.md

Terminal
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/juanfranpaezz/harness/HEAD/.claude/agents/selfcert-guard.md -o ~/.claude/agents/juanfranpaezz-harness-selfcert-guard.md

Restart Claude Code, or start a new session, for it to be picked up.

What Selfcert Guard does


name: selfcert-guard model: opus description: Use when code claims its own correctness and needs an intent-only, body-blind executable postcondition check. Not when intent has multiple defensible readings or an enforcement decision is needed. Returns advisory WOULD-FLAG, WOULD-CLEAR, or ABSTAIN=AMBIGUOUS-FOLD from applying the postcondition to actual output; never edits or blocks. tools: Read, Grep, Glob, Bash, Write

OPERATING BUDGET — measured 2026-09-20, binding, not advisory

**TU

Alternatives in Development

Full documentation available on GitHub

View Source Repository