CtrlK
BlogDocsLog inGet started
Tessl Logo

markusdowne/social-source-calibration

Calibrate research done on socially noisy web sources so agents do not mistake crowd mood for truth. Includes source-specific skills for Moltbook, Hacker News, Reddit, and Product Hunt.

92

1.07x
Quality

92%

Does it follow best practices?

Impact

100%

1.07x

Average score across 1 eval scenario

SecuritybySnyk

Passed

No known issues

Overview
Quality
Evals
Security
Files

SKILL.md

name:
social-source-calibration
description:
Route and calibrate research drawn from socially noisy web communities so agents do not mistake crowd mood for truth. Use when research notes, summaries, or quoted material come from Moltbook, Hacker News, Reddit, or Product Hunt and the job is to decide how much weight that material deserves, separate concrete weak signals from vibe/noise, or choose the right source-specific calibration skill before carrying findings forward.

Social Source Calibration

Use this tile as a source-routing and calibration layer for socially noisy research inputs.

Do not use it as a browsing skill, a fact-checking skill, or a license to treat crowd-generated material as proof.

What this tile is for

Use this tile when you already have gathered material from one of these sources and need to decide how much evidentiary weight it deserves:

  • Moltbook → weak-signal discovery with high noise and occasional concrete early reports
  • Hacker News → structural technical concerns mixed with anti-hype cynicism
  • Reddit → broad user pain and workaround language mixed with uneven expertise
  • Product Hunt → launch-day positioning and momentum mixed with supportive launch culture

Routing rule

Pick the source-specific skill that matches the material you are calibrating:

  • moltbook-source-calibration
  • hacker-news-source-calibration
  • reddit-source-calibration
  • product-hunt-source-calibration

If the material blends more than one source, calibrate each source separately first, then combine them conservatively.

Shared decision rule

Across all four sources:

  1. Separate concrete repeated detail from tone, popularity, or social heat.
  2. Decide whether the source is offering:
    • a follow-up-worthy weak signal
    • a single-post lead
    • low-weight noise
  3. State uncertainty explicitly.
  4. Gate the carry-forward decision: do not treat the claim as evidence until stronger-source verification is attempted.

Output shape

Prefer a compact output with:

  • classification
  • evidence strength
  • why
  • next check

Example:

  • classification: concrete report
  • evidence strength: follow-up-worthy weak signal
  • why: two separate sources describe the same concrete workflow failure, but both are still social reports rather than proof
  • next check: verify against issue trackers, docs, changelogs, or direct reproduction before carrying the claim forward

Read next

For source-specific calibration rules and examples, read docs/index.md, then the relevant subskill under skills/*/SKILL.md.

Guardrail

Never upgrade popularity, confidence, ranking, or repeated mood into proof. Social sources are for lead generation and calibration, not final truth.

SKILL.md

tile.json