skills.rest
sunny-kobe
sunny-kobeCOMMUNITY
Explore all skills by this author →
GitHub

sentinel-daily-report

Enforce quality gates for Project Sentinel daily report generation and publication.

21|5|Updated Jan 15, 2026
Search Tags:#evidence validation#close#rule engine#structured reporting#report quality#midday#AI narration guardrails
One-click install
npx skills add https://github.com/sunny-kobe/ai_sentiney --skill sentinel-daily-report
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: sentinel-daily-report
Source: https://github.com/sunny-kobe/ai_sentiney/tree/main/skills/sentinel-daily-report
Command: npx skills add https://github.com/sunny-kobe/ai_sentiney --skill sentinel-daily-report

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill prevents Project Sentinel midday or close narrations from being generated or published when trading-day, timestamp, evidence coverage, or underlying data completeness is insufficient.

Core Features & Use Cases

  • Quality gate (normal/degraded/blocked): Enforces checks for trading day, context_date alignment, portfolio availability, index/breadth presence, and message evidence before any AI narration or publishing.
  • Structured evidence-first output: Produces a structured daily report containing market overview, per-holding signals, deterministic operations, technical and news evidence, source labels, and data_timestamp.
  • Degraded publishing with strict constraints: Allows publishing only as a structured quick report (with explicit quality_status, data time, source labels, and rule-engine-driven operations), avoiding free-form model decisions when evidence or AI coverage is inadequate.
  • Verification support: Provides a test and dry-run command flow to validate report quality, structured rendering, and quality flow.

Quick Start

Run the midday mode in dry-run replay mode to verify report quality gates and structured rendering before any publishing steps.

Frequently Asked Questions about sentinel-daily-report

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I validate trading day and evidence coverage before generating A-share research reports?▼

To validate trading day and evidence coverage for A-share research reports, you must check trading day alignment, context_date, portfolio availability, and index/breadth presence. This ensures structured reporting only proceeds when underlying data completeness is sufficient.

What is a degraded mode for daily market report publishing and when is it needed?▼

Degraded mode for daily market report publishing allows a structured quick report to publish with explicit quality_status and source labels when evidence or AI coverage is inadequate. It prevents free-form model decisions by using rule-engine-driven operations.

How to run a dry-run replay to check structured rendering and quality gates for midday reports?▼

To run a dry-run replay for midday reports, execute the midday mode in dry-run replay mode. This validates report quality gates, structured rendering, and the normal/degraded/blocked decision workflow before any actual publishing steps occur.

Why does my AI narration guardrail block report generation when news evidence is missing?▼

AI narration guardrails block report generation when news evidence is missing to prevent publishing incomplete market summaries. The quality gate enforces evidence-first output, ensuring technical and news evidence presence before any AI narration proceeds.

Can I publish free-form model decisions if portfolio data is unavailable for the close report?▼

You cannot publish free-form model decisions if portfolio data is unavailable for the close report. The system enforces degraded publishing constraints, allowing only structured quick reports with explicit quality_status and data timestamps.

What specific data fields are required for structured evidence-first output in daily reports?▼

Structured evidence-first output in daily reports requires market overview, per-holding signals, deterministic operations, technical and news evidence, source labels, and data_timestamp. These fields ensure comprehensive evidence validation for A-share research.

Related Skills

Explore More Data & Analytics Skills→
quantskillsquantskills
42

market-daily-review

Generate A-share end-of-day market review reports from Pandadata.

Official
Advanced
mmbianco78mmbianco78

reports

Render Jinja2 and MJML templated reports from SQL data.

Community
Intermediate
PlaidmustachePlaidmustache

web-pentest

Identify web application attack surfaces with Hermes-guided recon and phase-based workflows.

Community
Advanced
hvkshetryhvkshetry
15

client-report

Generate client portfolio reports with performance, risk, and allocation data.

Community
Advanced