Skip to content
Back to skills

Terse Mode

ASecurity

Output-economy skill for research-heavy flows. Sacrifices prose polish for token efficiency in exploratory loops. VETO auto-off for code-review, security-engineer, compliance, debate consensus, audit artifacts.

  • 2 stars
  • 0 votes
  • 0 copies
  • 3 views
  • Added September 2, 2026
ai-agentscode-reviewfrontendsecuritydocumentation

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned September 2, 2026

npx -y skills add Canhada-Labs/ceo-orchestration --skill terse-mode --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Terse Mode?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Terse Mode
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/canhada-labs-terse-mode/badge)](https://www.skillsdirectory.com/skills/canhada-labs-terse-mode)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: Terse Mode
description: Output-economy skill for research-heavy flows. Sacrifices prose polish for token efficiency in exploratory loops. VETO auto-off for code-review, security-engineer, compliance, debate consensus, audit artifacts.
trigger: CEO activates via /terse on; auto-off for VETO roles; check_agent_spawn.py injects TERSE-DISABLED marker into spawn prompts of canonical-5.
# --- smart-loading fields (PLAN-083 Wave 0a sub-agent 0.7a) ---
domain: core
priority: 8
risk_class: low
stack: []
context_budget_tokens: 300
inactive_but_retained: false
repo_profile_binding:
  frontend: {active: true, priority: 8}
  engine: {active: true, priority: 8}
  fintech: {active: true, priority: 8}
  trading-readonly: {active: true, priority: 8}
  generic: {active: true, priority: 7}
activation_triggers:
  - {event: help-me-invoked, regex: "(?i)terse|concise|short"}
---

# Terse Mode

## Identity

When `/terse on` is active, CEO and non-VETO sub-agents produce
minimally-verbose outputs. Target: -40-60% output tokens in
exploratory research, bullet-list summaries, short factual answers.

## Rules

- **Tool first. Result first.** State the action/outcome in line 1.
- **Fragments OK** in: exploratory research, bullet lists, summaries,
  status updates, sanity checks.
- **Full sentences REQUIRED** in: code-review findings, security
  findings, debate consensus.md, verdict.md artifacts, audit reports,
  compliance write-ups, adopter-facing documentation.
- **Never truncate code** — code blocks remain verbatim.
- **Never drop numbers** — quantitative claims carry units + source.
- **Never use ellipsis to hide content** — if a finding matters,
  name it.

## VETO auto-off

Terse mode AUTO-DISABLES for these roles (hardcoded in
check_agent_spawn.py P0-TERSE-VETO guard):

- code-reviewer (merge VETO)
- security-engineer (auth/crypto VETO)
- qa-architect (when writing verdict.md)
- compliance-specialist (LGPD/GDPR findings)
- CEO writing consensus.md or verdict.md

For these roles, spawn prompt receives injected marker:
`## TERSE-MODE-DISABLED — VETO role requires full-prose rationale`.

## Toggle

- `/terse on` — activate for session
- `/terse off` — deactivate
- `/terse status` — print current state

## Measurement

Session cost delta is tracked via ceo-cost.py. Terse-mode sessions
emit `terse_mode_session` audit event with `mode_start_ts`,
`mode_end_ts`, `output_tokens_saved_estimate` fields.

## Adopter guidance

Terse mode IS NOT for:
- Production deploys (full audit trail needed)
- Customer-facing artifacts (quality over economy)
- Debate rounds (consensus quality depends on full reasoning)

Terse mode IS for:
- Research loops (scanning docs, grep-like tasks)
- Progress updates between milestones
- Internal logs where CEO is the only reader
- Repetitive sanity checks

Files in this skill

  • SKILL.md2.8 KB
  • SKILL.md.shadow.md2.3 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…