# Si Objector

> Use only when Selective Intelligence explicitly assigns a Guarded or Council review of a bounded Worker artifact and proof claim.

- Skill: `infotradescout/si-objector` (Agent Skill)
- Install (CLI): `npx skillmds@latest add infotradescout/si-objector`
- Raw SKILL.md: https://api.skillmd.com/api/skills/infotradescout/si-objector/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: infotradescout (https://skillmd.com/u/infotradescout)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/infotradescout/si-objector

---


# SI Sub-Skill: Objector

## What this skill does (plain language)
It checks one bounded Worker result like a careful safety inspector. Do not invoke it merely because work is persistent or agents are available.

## Inputs
- `si-worker` result packet
- Relevant evidence references (tests, commits, routes, files)

## Steps
1. Compare result to the lock plan.
2. Verify claims with evidence.
3. List missing work, weak evidence, drift, and scope skips.
4. Mark each finding with severity and exact place.
5. Recommend the smallest fix for each block.

## Output
Return:
- `findings` (numbered list with severity + evidence)
- `pass` / `sustained` / `partial` / `blocked`
- `next_skill`: `si-aligner`

## Non-negotiable rules
- No style-only policing.
- No invented success claims.
- No new scope that was not in the lock unless called out as a needed follow-up.
- Use plain language for all findings and why they matter.