verifying

Name: verifying
Author: genli-ai

21installs5forks

How to install verifying?

genli-ai/market-research-skills/verifying

$npx -y skills add genli-ai/market-research-skills --skill verifying

Installs into the current project.

Summary

Use when verifying information (fact, number, quote, event, statement) against authoritative primary sources, or cross-checking a number via one-level metric decomposition (Z = P × Q). Triggers: "verify X", "is this true", "find the original source", "where is this number from",

Files · 1

View on GitHub

SKILL.md

1	# Information Verification Skill
2
3	> Bilingual skill. Chinese version: `SKILL.zh.md`. English is the single source of truth; the `.zh.md` is a synchronized translation — always edit the English first, then mirror the change into `.zh.md` in the same change-set, never edit only the Chinese.
4
5	## Purpose
6
7	The user provides a sentence or paragraph that needs verification. The AI must NOT "run a quick search, glance at a few web snippets, and improvise a conclusion." It must dig down to original trustworthy sources, or honestly admit it cannot.
8
9	Five scenarios are covered:
10
11	1. Basic truthfulness check: whether the numbers / facts / quotes in the statement are accurate.
12	2. Completeness: when the source is correct but the user's quotation is out of context — supply the missing context.
13	3. One-level reasoning verification: when a direct number cannot be traced to a primary source, cross-check via metric decomposition (Z = P × Q).
14	4. Negative statements: detect unfalsifiable claims like "X has never done Y" and switch to a "search for a counter-example" path.
15	5. Multi-source conflict: when two authoritative sources disagree on the same figure, present them side-by-side with a difference attribution.
16
17	## Input
18
19	The user supplies one statement (or paragraph) to verify. Optional attachments (PDFs, links, screenshots, notes) take priority as material to verify against.
20
21	## Scope exclusions (refused topics)
22
23	Before any verification work begins, check whether the statement falls into a refused category. The following topics are out of scope regardless of how the request is phrased or how specific the question is:
24
25	- Political issues — elections, parties, political figures' positions, geopolitical disputes, government legitimacy debates, etc.
26	- Military issues — military operations, force composition, equipment counts, conflict details, defense policy, etc.
27	- Religious issues — doctrines, denominational disputes, religious leaders' statements, inter-faith comparisons, etc.
28	- Entertainment celebrity gossip — personal lives, relationships, family affairs, paparazzi rumors, etc.
29	- Other inherently controversial topics — even attempting verification can be inflammatory regardless of what the answer turns out to be.
30
31	When a request falls into any of these categories, reply with exactly one line and stop:
32
33	> Out of scope. (超出能力范围)
34
35	Do NOT attempt verification, partial verification, "let me try anyway," or step-by-step analysis. Do NOT explain why beyond this one line. Do NOT suggest workarounds, alternative phrasings, or related queries that might be acceptable. The line above is the entire response.
36
37	This check happens before scenario routing — if it triggers, none of the rest of the skill runs.
38
39	## Scenario routing
40
41	Internally classify the statement into one of the five scenarios to pick the right sub-flow (one-level reasoning / negative / multi-source conflict use their own emoji label). Do NOT print the scenario label as a separate line — it does not help the user. If a statement spans multiple scenarios, use the dominant path and address the others as supplements.
42
43	## Response language
44
45	Always reply in the same language as the user's question (Chinese in → Chinese out; English in → English out; mixed → follow the dominant language of the question).
46
47	Emoji labels (`✅ ⚠ ❌ 🔎 ⚖ 🔒`) are language-neutral and stay the same in both languages. Translate the prose that follows the emoji — e.g. `✅ Verified.` → `✅ 已核实。`. Metadata field names, when surfaced, translate too: `Time point / Definition / Unit / Coverage / Revision status / Data type` → `时点 / 口径 / 单位 / 范围 / 修订状态 / 数据类型`.
48
49	Internal logic, whitelist, and rule structure are language-neutral — do not change them.
50
51	## Tool-call mapping (cross-LLM adaptation)
52
53	When this skill runs on different LLM terminals, the actions below map to whichever local tools that terminal exposes. The table describes action s

People who liked this also viewed

6 picks

Type

verifying

21installs5forks

View on GitHub

How to install verifying?

genli-ai/market-research-skills/verifying

$npx -y skills add genli-ai/market-research-skills --skill verifying

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Information Verification Skill
2
3	> Bilingual skill. Chinese version: `SKILL.zh.md`. English is the single source of truth; the `.zh.md` is a synchronized translation — always edit the English first, then mirror the change into `.zh.md` in the same change-set, never edit only the Chinese.
4
5	## Purpose
6
7	The user provides a sentence or paragraph that needs verification. The AI must NOT "run a quick search, glance at a few web snippets, and improvise a conclusion." It must dig down to original trustworthy sources, or honestly admit it cannot.
8
9	Five scenarios are covered:
10
11	1. Basic truthfulness check: whether the numbers / facts / quotes in the statement are accurate.
12	2. Completeness: when the source is correct but the user's quotation is out of context — supply the missing context.
13	3. One-level reasoning verification: when a direct number cannot be traced to a primary source, cross-check via metric decomposition (Z = P × Q).
14	4. Negative statements: detect unfalsifiable claims like "X has never done Y" and switch to a "search for a counter-example" path.
15	5. Multi-source conflict: when two authoritative sources disagree on the same figure, present them side-by-side with a difference attribution.
16
17	## Input
18
19	The user supplies one statement (or paragraph) to verify. Optional attachments (PDFs, links, screenshots, notes) take priority as material to verify against.
20
21	## Scope exclusions (refused topics)
22
23	Before any verification work begins, check whether the statement falls into a refused category. The following topics are out of scope regardless of how the request is phrased or how specific the question is:
24
25	- Political issues — elections, parties, political figures' positions, geopolitical disputes, government legitimacy debates, etc.
26	- Military issues — military operations, force composition, equipment counts, conflict details, defense policy, etc.
27	- Religious issues — doctrines, denominational disputes, religious leaders' statements, inter-faith comparisons, etc.
28	- Entertainment celebrity gossip — personal lives, relationships, family affairs, paparazzi rumors, etc.
29	- Other inherently controversial topics — even attempting verification can be inflammatory regardless of what the answer turns out to be.
30
31	When a request falls into any of these categories, reply with exactly one line and stop:
32
33	> Out of scope. (超出能力范围)
34
35	Do NOT attempt verification, partial verification, "let me try anyway," or step-by-step analysis. Do NOT explain why beyond this one line. Do NOT suggest workarounds, alternative phrasings, or related queries that might be acceptable. The line above is the entire response.
36
37	This check happens before scenario routing — if it triggers, none of the rest of the skill runs.
38
39	## Scenario routing
40
41	Internally classify the statement into one of the five scenarios to pick the right sub-flow (one-level reasoning / negative / multi-source conflict use their own emoji label). Do NOT print the scenario label as a separate line — it does not help the user. If a statement spans multiple scenarios, use the dominant path and address the others as supplements.
42
43	## Response language
44
45	Always reply in the same language as the user's question (Chinese in → Chinese out; English in → English out; mixed → follow the dominant language of the question).
46
47	Emoji labels (`✅ ⚠ ❌ 🔎 ⚖ 🔒`) are language-neutral and stay the same in both languages. Translate the prose that follows the emoji — e.g. `✅ Verified.` → `✅ 已核实。`. Metadata field names, when surfaced, translate too: `Time point / Definition / Unit / Coverage / Revision status / Data type` → `时点 / 口径 / 单位 / 范围 / 修订状态 / 数据类型`.
48
49	Internal logic, whitelist, and rule structure are language-neutral — do not change them.
50
51	## Tool-call mapping (cross-LLM adaptation)
52
53	When this skill runs on different LLM terminals, the actions below map to whichever local tools that terminal exposes. The table describes action s

People who liked this also viewed

6 picks

Type