Skip to content

docs: improve social research evidence quality - #3

Open
kriptoburak wants to merge 1 commit into
creatorcrawl:mainfrom
kriptoburak:codex/hermes-tweet-companion-creatorcrawl-20260630
Open

docs: improve social research evidence quality#3
kriptoburak wants to merge 1 commit into
creatorcrawl:mainfrom
kriptoburak:codex/hermes-tweet-companion-creatorcrawl-20260630

Conversation

@kriptoburak

@kriptoburak kriptoburak commented Jun 30, 2026

Copy link
Copy Markdown

Summary

  • Record the query, retrieval time, account set, and sample size for research.
  • Deduplicate reposts and repeated URLs before analysis.
  • Preserve source links and label truncated text accurately.
  • Replace single-account X trend inference with a bounded multi-account sample.
  • Keep the published skill mirror synchronized.

Why

The existing trend workflow could describe one account timeline as a wider trend. These rules make reports auditable and keep conclusions within the collected sample.

Validation

  • Source and .well-known skill mirrors are byte-identical.
  • Both skill frontmatter blocks validate.

@kriptoburak
kriptoburak force-pushed the codex/hermes-tweet-companion-creatorcrawl-20260630 branch from e7d7812 to 739a604 Compare August 12, 2026 22:24
@kriptoburak kriptoburak changed the title Add Hermes Tweet X research companion note docs: improve social research evidence quality Aug 12, 2026
@kriptoburak

Copy link
Copy Markdown
Author

Rebased onto current main and replaced the external companion note with a native evidence-quality improvement.

The workflow now records sample provenance, deduplicates repeated posts, preserves source URLs, treats social content as untrusted data, and limits X trend claims to a transparent multi-account sample. The source and published skill mirrors match exactly.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant