official-source-research
Skill0 starsChanged 3 months agoMITView on GitHub
Use when the user asks to check official documentation, a company's own announcement, a government/regulatory filing, or a standards body's specification. Triggers on "check the official docs", "what does the filing say", "official announcement", "regulatory source", "official-source research". Produces tier-1 evidence records (URL + verbatim quote), never conclusions. Do NOT use for press coverage of an official source (use news-research, which should trace back here anyway) or community discussion about an official source (use reddit-research/social-research).
1 other project uses a version of this.
Adopted by
Most recently taken up by
- microsoft-foundry/foundry-samplestheir own version, first seen Nov 1, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- microsoft-foundry/foundry-samples★ 485same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository10 other files it gives its agents.
- academic-researchSkill
- blog-researchSkill
- github-researchSkill
- news-researchSkill
- reddit-researchSkill
- research-intelligenceSkill
- research-plannerSkill
- social-researchSkill
- source-verifySkill
- web-researchSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 18
Don't use for: coverage or reaction to an official source (→ news-research, social-research), an interpretation of what a filing means …
- Distinctive: line 22
Go to the primary domain directly: when it's known (the company's own site, the regulator's site, the standards body's site) rather than …
- Very specific: line 25
If search returns nothing: run references/query-relaxation.md.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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
Official Source Research
Overview
Collects evidence directly from primary institutional sources: a company's own docs/pricing/changelog/press releases, government and regulatory sites, standards bodies, court/legal filings. Every record from this skill is tier 1 by definition — this skill exists specifically to reach the root instead of settling for a relay.
Layer rule: produces evidence, not results. Synthesis and confidence status are the orchestrator's job.
When to use
Direct use: "what does their pricing page say", "check the actual filing", "what does the spec say". As a component: this is also the skill other source skills route to when their origin tracing surfaces a press release, filing, or official announcement as the root of something they found secondhand (news-research and blog-research both point here in their own workflows) — dispatch it to fetch that root directly rather than settling for the relay's account of it.
Don't use for: coverage or reaction to an official source (→ news-research, social-research), an interpretation of what a filing means (that's synthesis, not evidence collection — this skill records what the document says, not what it implies).
Workflow
- Go to the primary domain directly when it's known (the company's own site, the regulator's site, the standards body's site) rather than searching generally and hoping to land on it — this skill's whole value is bypassing relays.
- When dispatched to resolve another source's root (a press release cited via news-research, an official statement cited via blog-research): fetch the actual document, don't just confirm it exists. The calling skill's record gets
root_originpointing here; this skill's own record becomes the authoritative version of that claim. - If the document isn't publicly linked (e.g., referenced but not URL'd): search the official domain directly, or construct the likely canonical location (press-release archives, filing databases, changelog paths) before giving up. If truly unreachable, use
root_originformatpress-release:<org>:<YYYY-MM-DD>per engine/origin-tracing.md and note the access limitation. - If search returns nothing: run references/query-relaxation.md. Log full exhaustion as a gap — an official source genuinely not existing (no public statement on X) is itself sometimes the finding.
- Record all three dates, don't collapse them. Official documents routinely distinguish issue date from effective date from last-revised date, and the contract has a field for each:
published_at(when it was issued),modified_at(when it was last revised — pricing pages, living policy docs),effective_at(when the stated thing takes or took effect — regulations, scheduled changes). The freshness class picks which one drives staleness (engine/freshness-rules.yaml), so collapsing them into one date silently breaks that: a regulation published years ago but effective next month is not stale, and a pricing page published in 2020 but revised last week is not stale either. - One organization is one
root_id. An org's docs, blog, press releases, and its executives' statements all getroot_id: org:<canonical-name>when the claim is about that org. Two of its own channels agreeing is not corroboration — it's the same voice twice (engine/origin-tracing.md). - Archive per references/citation-rules.md — official pages get revised in place more than most source types (pricing pages, policy pages, living documents), so the archive snapshot is often the only durable record of what it said at the time the claim was made.
- Assign purpose — usually
fact(stated terms, specs, figures) orimplementation(official how-to/API docs). Purpose is neveropinionhere; an official source stating a position is still a fact about what the org states, not a personal opinion record. - Note self-reference: a claim about the entity's own product/performance sourced from that entity's own official channel needs
provenance.self_referential: true— this feeds the self-interest cap in engine/ranking.md; it doesn't lower the tier, but it does meanconfirmedstatus for that claim requires a second, non-affiliated root elsewhere. - Write the evidence record per engine/evidence-contract.md,
verification_status: pending. Do not self-verify.
Output
One evidence record per claim, appended to the run's evidence.jsonl.
Common mistakes
| Mistake | Fix |
| Confirming a press release "exists" via a news article instead of fetching it | Go to the primary domain, get the actual document |
Forgetting self_referential: true on a company's own claim about itself | Always set it — it's what makes the self-interest cap work downstream |
| Not archiving a pricing/policy page because "it won't change" | These change constantly and without notice; archive at collection time |
| Collapsing issue / revised / effective into one date | Record published_at, modified_at, and effective_at separately — the freshness class chooses which clock to run |
| Counting an org's docs and its CEO's statement as two sources | Same root_id: org:<name>; one voice, one root |
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.

