skills.rest
line7works
line7worksOFFICIAL
Explore all skills by this author →
WebsiteGitHub

delivery-probe

Echoes sentinel lines from its own body to verify skill delivery.

Updated Jun 2, 2026
Search Tags:#testing#fixture#sentinel#delivery verification#prompt assembly#skill harness
One-click install
npx skills add https://github.com/line7works/tony-skills --skill delivery-probe-line7works
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: delivery-probe
Source: https://github.com/line7works/tony-skills/tree/main/plugins/recheck-v2/setups/_fixtures/delivery-probe/skills/delivery-probe
Command: npx skills add https://github.com/line7works/tony-skills --skill delivery-probe-line7works

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? It verifies whether a skill harness actually delivers the full body of a SKILL.md file into the model's context, exposing truncation or catalog budget cuts during a skills v2 pilot. ## Core Features & Use Cases - Sentinel Echo Test: Replies with every SENTINEL S01–S24 line found in its body, in order, followed by END-OF-PROBE. - Delivery Measurement: Uses a long padded body so any harness-imposed prompt budget or truncation becomes visible as missing sentinels. - Use Case: A developer testing a new skill-loading harness invokes the probe and checks whether all 24 sentinel lines come back, confirming complete body delivery. ## Quick Start Ask the assistant to run the delivery probe and check that the reply lists sentinels S01 through S24 followed by END-OF-PROBE.

Frequently Asked Questions about delivery-probe

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

FAQPage Schema
How do I test whether a skill body is fully delivered to the model?▼

Invoke the delivery probe and inspect the reply. It echoes every SENTINEL line S01 through S24 it can see, followed by END-OF-PROBE, so any missing sentinel indicates truncation by the harness.

What is a sentinel-based delivery probe for AI skills?▼

It is a test fixture skill whose body contains numbered sentinel lines embedded in padding text. The model must repeat them verbatim, making it easy to detect whether catalog or prompt budgets cut off part of the body.

When should I use the delivery probe fixture?▼

Use it only when validating a skill-loading pipeline, such as during the skills v2 pilot. It is a fixture under recheck-v2 test setups and explicitly should never be used for any other request.

Why does the delivery probe reply miss some sentinel lines?▼

Missing sentinels mean the harness truncated the skill body before it reached the model, typically due to a catalog or prompt size budget. The probe body is intentionally long to cross such budgets and expose them.

Does the delivery probe read files or run scripts?▼

No. Its instructions forbid reading any file; only the body delivered into context counts. It contains no scripts, references, or assets directories.

Related Skills

Explore More Software Engineering Skills→
harnessprotocolharnessprotocol
9

harness-docs

Answers questions about installing, configuring, and understanding harness-kit plugins and workflows.

Official
Basic
wan-huiyanwan-huiyan
3

cjk-structured-llm-output-truncates-json-needs-2x-tokens

Diagnose and fix truncated JSON from long CJK LLM outputs by budgeting tokens and salvaging partial objects.

Community
Intermediate
compozycompozy
2.7k

portable-skill-12

Provides a minimal portable skill definition for testing skill loading.

Official
Basic
skaile-aiskaile-ai

test-count-artifacts

Count regular files under workspace/test and write the count to count.txt.

Official
Basic