page-report
Skill51 starsChanged 6 months agoMITView on GitHub
Shared report-page contract for Agentic SEO modules that render editable, human-first Markdown reports in the Web Companion.
In this repository22 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- agentic-seoSkill
- autoresearchSkill
- backlink-analysisSkill
- brain-keeperSkill
- competitive-analysisSkill
- content-importSkill
- content-seoSkill
- data-setupSkill
- eeatSkill
- internal-linksSkill
- keyword-researchSkill
- project-initSkill
- seo-analysisSkill
- seo-skills-creatorSkill
- seo-tools-creatorSkill
- serp-extractSkill
- spec-drivenSkill
- startSkill
- technical-seoSkill
- topic-clusterSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
13 rules. About 0 are common, 5 are distinctive, and 8 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
8
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 11
Use this skill whenever an Agentic SEO workflow generates a report page for the Web Companion.
- Very specific: line 15
Use page-report from report-producing skills and runtime code when creating or reviewing report.md for Technical SEO, Internal Links, SEO …
- Distinctive: line 17
Do not use this skill for public content drafts, approval handoffs, inline chat answers, raw provider captures, or authorial Brain updates …
- Very specific: line 24
Do not write # <title> in the Markdown body.
- Very specific: line 25
Use the project language from project/.agentic-seo/project.json.language unless a command has an explicit language override.
- Very specific: line 30
Before declaring a manual report done, run npm run validate:reports -- --module=<module> --slug=<run-slug> or use …
- Very specific: line 31
Never paste raw provider JSON, object arrays, [object Object], internal IDs, or snake_case evidence into the visual body.
- Very specific: line 33
Always return report_md and browser_prompt: { recommended: true, message: "Posso abrir o Web Companion para você ver a análise?", …
- Very specific: line 40
Write the module's authoritative machine-readable artifact first.
- Distinctive: line 44
Start with an executive reading.
- Distinctive: line 48
Before returning the result, verify: canonical path, frontmatter, no H1 body title, no HTML default, YAML visual fences, no raw JSON/object …
- Very specific: line 80
Use only supported report_type values: technical-seo, internal-links, seo-analysis, keyword-research, serp-extract, backlink-analysis, …
- Distinctive: line 88
No visible JSON/object dumps or duplicated title appear in the report body.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Page Report
Use this skill whenever an Agentic SEO workflow generates a report page for the Web Companion. It is a shared contract, not a standalone SEO module.
When To Use
Use page-report from report-producing skills and runtime code when creating or reviewing report.md for Technical SEO, Internal Links, SEO Analysis, Keyword Research, SERP Extract, Backlink Analysis, Topic Cluster, E-E-A-T, or Competitive Analysis.
Do not use this skill for public content drafts, approval handoffs, inline chat answers, raw provider captures, or authorial Brain updates that are not report pages.
Critical Points
- The canonical report path is
project/analyses/<module>/<run-slug>/report.md. - The report is the editable human presentation layer. Raw evidence, normalized YAML/JSON, logs, and operational artifacts stay in
sources/,audits/,keywords/,clusters/,eeat/, orworkbench/. - Frontmatter must include
title,slug,report_type,generated_at,status,source_artifact, andsummary; includescoreonly when the module has a real score. - Do not write
# <title>in the Markdown body. The Companion renders the title from frontmatter. - Use the project language from
project/.agentic-seo/project.json.languageunless a command has an explicit language override. Complete copy is required forpt-BRanden. - New visual fences must be YAML with
version: 1:agentic-kpis,agentic-chart, andagentic-table. JSON fences are legacy read compatibility only. - Tables must use stable lower-snake ASCII
keyvalues. User-facinglabelvalues may be translated or edited without changing keys. - Calculation tables must mark columns with
role: weight,role: points, androle: lossso score recalculation survives label edits. - Each report-producing module has
skills/<module>/contract.yamlandtemplates/analyses/<module>/report-skeleton.md; keep required H2 sections and table keys from the skeleton. - Before declaring a manual report done, run
npm run validate:reports -- --module=<module> --slug=<run-slug>or usescripts/lib/report-writer.mjs, which refuses invalid reports before writing. - Never paste raw provider JSON, object arrays,
[object Object], internal IDs, or snake_case evidence into the visual body. Convert evidence into natural language and readable tables. - Format numbers per the project language. The canonical source is
shared/locale.mjs(getProjectLanguage,formatNumber,formatPercent,canonicalKeyword). Skills either ship locale-formatted strings inagentic-table/agentic-kpiscells, or leave numbers raw for the Companion (useI18n().formatNumber/formatPercent) to format on render. Percent-bearing columns use the_pctkey suffix on the 0..100 scale. - Always return
report_mdandbrowser_prompt: { recommended: true, message: "Posso abrir o Web Companion para você ver a análise?", open_with: "project-browser" }. - Register the generated report in
project/brain/log.mdastype: decisionunless a more specific log type is required.
Framework
1. Build The Source Artifact
Write the module's authoritative machine-readable artifact first. It may be YAML or JSON, but it must remain outside project/analyses/ and be referenced by frontmatter source_artifact.
2. Render The Human Page
Start with an executive reading. Then add module-specific sections, visual blocks, and human-readable appendices. Keep technical depth, but translate it into prose, friendly labels, and compact evidence.
3. Validate The Contract
Before returning the result, verify: canonical path, frontmatter, no H1 body title, no HTML default, YAML visual fences, no raw JSON/object dumps, readable evidence, source separation, and log entry.
4. Review The Result
For QA, generate one report per module with fixtures/offline data. A report is satisfactory only when it scores at least 4/5 for clarity, evidence separation, visual usefulness, actionability, and localization, with no category below 3/5.
Output Format
The report Markdown must have this shape:
---
title: "Human report title"
slug: "stable-run-slug"
report_type: "technical-seo"
generated_at: "2026-05-23T00:00:00.000Z"
status: "ready"
source_artifact: "audits/example/report.yaml"
summary: "One-sentence executive summary."
score: "92"
---
Executive summary prose without a duplicate H1.
```agentic-kpis
version: 1
items:
- label: Score
value: 92/100
```Use only supported report_type values: technical-seo, internal-links, seo-analysis, keyword-research, serp-extract, backlink-analysis, topic-cluster, eeat, and competitive-analysis.
Done Criteria
report.mdexists under the canonical module path and is returned asreport_md.source_artifactpoints to the normalized technical source.- The Companion can list, open, edit, save, and roundtrip the page.
- Visual fences parse as YAML
version: 1. - No visible JSON/object dumps or duplicated title appear in the report body.
- The report is registered in
brain/log.md.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

