Skip to content
Back to skills

127 Cross Afd97cff

ASecurity

Cross-reference multiple sources for verification.

  • 9 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added October 11, 2026
researchgo

Security analysis

A100/100

Scanned October 11, 2026

npx -y skills add tools-only/X-Skills --skill 127-cross_afd97cff --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of 127 Cross Afd97cff?

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

Security grade badge for 127 Cross Afd97cff
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/tools-only-127-cross-afd97cff/badge)](https://www.skillsdirectory.com/skills/tools-only-127-cross-afd97cff)

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

SKILL.md
# Triangulating Mode

Cross-reference multiple sources for verification.

## When to Use

- Claim needs validation
- Single source insufficient
- Stakes are high
- Information seems surprising
- Source reliability uncertain

## Mental Model

**Navigation by triangulation:** Fix position from multiple bearings. Where they converge = likely truth.

## When Triangulation is Required

### Stakes-Based Threshold

| Stakes | Requirement |
|--------|-------------|
| Critical | 3+ independent sources |
| High | 2+ independent sources |
| Medium | 1 source + verification check |
| Low | Single source acceptable |

### Mandatory Triggers

- Claims that would drive major decisions
- Information that seems too good/bad to be true
- Surprising or counterintuitive claims
- Claims contradicting prior beliefs
- Claims from unknown/unreliable sources

## Triangulation Patterns

### 3-Independent
Three unrelated sources agreeing.

**Confidence boost:** High

**Use when:** Critical decisions, surprising claims.

### Primary-Secondary
Original source plus analyst interpretation.

**Confidence boost:** Medium

**Use when:** Need both data and context.

### Multi-Method
Same question, different methodologies.

**Confidence boost:** High

**Use when:** Methodology could bias results.

### Temporal
Same source over time.

**Confidence boost:** Medium

**Use when:** Checking for consistency over time.

## Execution

### Source Independence

**Independent sources:**
- Different organizations
- Different methodologies
- No shared data source
- No common author/analyst

**NOT independent (common mistakes):**
- Same original source, different publishers
- Derived from same data
- Same author in different venues
- Press release repackaged by multiple outlets

### Process

1. **Identify claim** — State the claim clearly, note original source, assess initial confidence
2. **Gather sources** — Find 3+ sources, verify independence, note source type and reliability
3. **Assess agreement** — Do sources agree? Where do they differ?
4. **Analyze conflicts** — For disagreements, list possible explanations, assess which is most likely
5. **Form conclusion** — State verdict with confidence, acknowledge caveats

### Confidence Adjustment

| Agreement Level | Confidence Adjustment |
|-----------------|----------------------|
| Full (all sources agree) | +20-30% |
| Majority (most agree) | +10-20% |
| Mixed (no clear pattern) | No change |
| Contradictory (sources disagree) | -10-20% |

## Output Format

```markdown
## Triangulating: [Claim]

**Original claim:** [Statement]
**Original source:** [Where it came from]
**Initial confidence:** [X%]
**Stakes:** [Critical/High/Medium/Low]

### Sources Examined

**Source 1: [Name]**
- Type: [Primary/Secondary/Tertiary]
- Reliability: [High/Medium/Low]
- Finding: [What this source says]
- Supports claim: [Yes/No/Partial]
- Independent from others: [Yes/No — why]

**Source 2: [Name]**
- Type: [Primary/Secondary/Tertiary]
- Reliability: [High/Medium/Low]
- Finding: [What this source says]
- Supports claim: [Yes/No/Partial]
- Independent from others: [Yes/No — why]

**Source 3: [Name]**
- Type: [Primary/Secondary/Tertiary]
- Reliability: [High/Medium/Low]
- Finding: [What this source says]
- Supports claim: [Yes/No/Partial]

### Agreement Analysis

**Agreement level:** [Full/Majority/Mixed/Contradictory]

**Where sources agree:**
- [Agreement point]

**Where sources differ:**
- [Disagreement]: [Source A says X, Source B says Y]
  - Possible explanations: [List]
  - Most likely: [Explanation]

### Conclusion

**Verdict:** [Validated / Partially validated / Uncertain / Invalidated]

**Refined claim:** [If claim needs modification based on findings]

**Confidence:** [X%] — [Why this confidence level]

**Evidence strength:** [Strong/Moderate/Weak]

**Caveats:**
- [Important caveat or limitation]

### Next

**If validated:** [What to do]
**If uncertain:** [What additional verification needed]
**If invalidated:** [What to do]
```

## Quality Gates

| Gate | Requirement |
|------|-------------|
| ≥3 sources attempted | For high-stakes claims |
| Independence verified | Not echoes of same source |
| Conflicts analyzed | Disagreements investigated |
| Confidence adjusted | Based on agreement level |
| Caveats documented | What might still be wrong |

## Anti-Patterns

| Avoid | Problem | Do Instead |
|-------|---------|------------|
| Non-independent sources | False corroboration | Verify true independence |
| Counting echoes | Same info, multiple outlets | Trace to original source |
| Dismissing disagreement | Miss truth | Investigate conflicts |
| Confirmation bias | Favor agreeing sources | Weight by reliability |
| Premature conclusion | Insufficient verification | Require threshold sources |

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…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.