hashgraph-online/awesome-codex-plugins / plugins/alpoxdev/hypercore/skills/hyper-research
hyper-research
Verified[Hyper] Use this skill when the user asks for multi-source, cited research - fact-finding, comparisons, market or trend analysis, or evidence-backed recommendations across live web, official docs, GitHub, and local repo sources - saved as a markdown report. Do not use for one-source lookups, rewriting, or implementation.
| File | Kind | Size |
|---|---|---|
| SKILL.ko.md | markdown | 20.6 KB |
| SKILL.md | skill md | 17.6 KB |
| assets/evals/hyper-research-cases.jsonl | other | 23.4 KB |
| references/channel-selection.ko.md | markdown | 5.7 KB |
| references/channel-selection.md | markdown | 5.2 KB |
| references/journal-format.ko.md | markdown | 7.6 KB |
| references/journal-format.md | markdown | 7.0 KB |
| references/latex-report.ko.md | markdown | 7.0 KB |
| references/latex-report.md | markdown | 6.4 KB |
| references/report-template.ko.md | markdown | 5.3 KB |
| references/report-template.md | markdown | 5.0 KB |
| rules/claim-verification.ko.md | markdown | 9.8 KB |
| rules/claim-verification.md | markdown | 8.5 KB |
| rules/expansion-and-convergence.ko.md | markdown | 6.7 KB |
| rules/expansion-and-convergence.md | markdown | 5.4 KB |
| rules/parallel-research.ko.md | markdown | 6.2 KB |
| rules/parallel-research.md | markdown | 5.3 KB |
| rules/research-workflow.ko.md | markdown | 8.3 KB |
| rules/research-workflow.md | markdown | 7.2 KB |
| rules/validation.ko.md | markdown | 11.6 KB |
| rules/validation.md | markdown | 10.1 KB |
| scripts/report-state.shexecutable | script | 16.5 KB |
| scripts/validate-report.shexecutable | script | 16.2 KB |
references/journal-format.md
View on GitHub# Journal Format
Use this reference before opening the journal directory for a default, deep, or parallel run. `--quick` keeps the ledger and query log inside the report instead.
## Contents
- Directory layout
- `brief.md`
- `sources.md`
- `queries.md`
- `expansion.md`
- `claims.md`
- `debate.md`
- `verify-<slug>.md`
- Writing rules
## Directory Layout
```text
.hyper/research/
├── [NN].slug.md # the saved report
└── [NN].slug/ # the journal, same stem
├── brief.md
├── sources.md
├── queries.md
├── expansion.md
├── claims.md
├── debate.md # high-risk claim attacks, when any
├── report-state.json # repair passes and the outcome manifest
├── outcome.json # written by report-state.sh finish
└── verify-<slug>.md # one per execution check, when any
```
The leader writes every file. Lanes return text; they never write here.
## `brief.md`
```markdown
# Brief: <topic>
Run started: YYYY-MM-DD HH:MM | Depth: default | Parallel: no
## Analysis
Core question:
Axes:
- A1 <axis> - <what, where, why>
Expected truths:
- E1 <statement> - status: true | violated | unknown
Entity: <canonical name, first-party URL, confidence>
Codebase relevant: | External: | Execution checks likely: | Precision demand:
## Fidelity
as of: latest at run start, YYYY-MM-DD (answered_by: default)
source floor: standard (answered_by: default)
derived numbers: allowed (answered_by: default)
asked: <fields asked, or none>
## Steering
- <exact user wording> - <wave it took effect>
## Deliverable
format: md (answered_by: default | user | request)
path: .hyper/research/[NN].slug.md
```
The `Deliverable` promise names what the run promises to deliver: the format (`md` by default), its path, and the `answered_by` value that chose the format. When the user asked for a different format, that answer is folded into the `Fidelity` block like the other fields; otherwise `answered_by: default` stands.
## `sources.md`
One row per source the moment it is opened. `S<n>` ids are stable for the whole run and are what the report cites.
```markdown
| # | Source | URL/path | Publisher | Published/updated | Accessed | Version/freshness | Channel | Grade | Role | Relevant claim | Used? | Caveat |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| S1 | | | | | | | | S/A/B/C | lead/primary-evidence/supporting/conflict/rejected | | yes/no | |
```
- Grade: S for primary, official, standard, or first-party; A for independent methodology or major institutions; B for practitioner or vendor material; C for promotional, unattributed, outdated, or lead-only.
- A row written from a snippet or a lane's summary is marked `unverified` in `Caveat` until the artifact is opened.
- Under each high-risk row, keep the exact quoted span or line the claim rests on.
- A source reached through a blocked-page route records the route and its provenance label in `Channel`, a snapshot timestamp in `Version/freshness`, and the access report path in `Caveat` (`references/channel-selection.md`, Blocked Sources). An unreachable source keeps its row with `Used?` `no` and the blocking signal in `Caveat`.
## `queries.md`
```markdown
| # | Query / command | Channel | Axis or lead | Why this query | Result | Follow-up |
|---|---|---|---|---|---|---|
```
No identical or semantically equivalent query twice. Counter-searches for the claim gate are logged here with `counter:` in the `Why` column.
## `expansion.md`
```markdown
## Wave 1
Searches or lanes: <list>
Returns:
- <lane or axis>: <digest>
## EXPAND (verbatim)
Leads opened: L1 <lead>, L2 <lead>
Leads closed: L0 duplicate of ... | dead end because ...
## Excursions
| id | parent claim or axis | trigger | budget | depth | exit rule | changed in the answer |
|---|---|---|---|---|---|---|
## Run state
converged | exhausted (cap, open leads, last result, extension plan) | blocked (missing access)
```
## `claims.md`
```markdown
## verified-claims
- C1: <statement>
## Claim table
| claim_id | statement | risk | sources | domains | primary | counter-search | dates (published / accessed / valid as of) | lineage | status |
|---|---|---|---|---|---|---|---|---|---|
| C1 | | high | S1, S4 | 2 | S1 | Q7: none stronger | | MEASURED | supported |
## Unresolved
- C5: <statement> - <what evidence is missing>
## Refuted
- C6: <statement> - <the counter-evidence, with S ids>
```
Status is `supported`, `partial`, `refuted`, or `unresolved`. Only `## verified-claims` rows may be asserted as high-risk statements in the report.
## `debate.md`
One row per attack on a high-risk claim, in the slot order `rules/claim-verification.md` §7 fixes:
```markdown
| claim id | attack | attacker perspective | defense | verdict (holds, weakened, withdrawn) | change to the claim |
|---|---|---|---|---|---|
| C1 | <the objection> | evidence quality | <the rebuttal> | holds | |
| C1 | <the objection> | source independence | <the rebuttal> | weakened | narrowed to the surveyed segment |
```
Two attacks count only when their `attacker perspective` values differ. A high-risk claim with fewer than two rows here stays `unverified` and out of `## verified-claims`, and the report's `## Debate record` annex lists each claim's verdict.
## `report-state.json` and `outcome.json`
`scripts/report-state.sh` owns these two files; never hand-edit them. `repair` appends each pass and its defect codes to `report-state.json`. `finish` refuses delivery while a gate is unset, a blocker remains, or the report or ledger is empty, and writes `outcome.json` beside the state file only when it accepts the run. `briefing` prints the counted facts from `outcome.json` for the closing effort line.
## `verify-<slug>.md`
```markdown
# Verify: <claim>
Claim id: C3 | Source: S2 | Contradiction: S5
Environment: <OS, runtime, dependency versions>
## Code
## Output
## Verdict
CONFIRMED | REFUTED | PARTIAL - <grounded in the output above>
Cleanup: <temporary directory removed, path>
```
## Writing Rules
- Append when an event happens, never in an end-of-run batch.
- Never rewrite a past row to change what was observed; add a new row and mark the old one `rejected` or `conflict` with the reason.
- Keep absolute dates everywhere; no relative dates.
- Store distilled claims with their source and date, never raw retrieved page text.
## Sources
> Checked 2026-11-03.
| Claim | Source |
|---|---|
| A leader-owned session journal written in real time, with separate source, expansion, claim, and verification files, adapted from the `ulw-research` skill | <https://github.com/code-yeongyu/oh-my-openagent/tree/0d80470e4201925727e88afbafcdf24d7801a588> |
| Ledger columns, grades, roles, the query log, and agent rows staying unverified until opened | `instructions/sourcing/references/source-ledger.md` |
| Never persist raw retrieved text into durable state | `instructions/sourcing/references/retrieval-safety.md` |
### Evidence grade
`LOCAL` - this file states this package's own journal layout. File names and columns are this package's choices; the upstream skill uses a different file set.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-changelog-updatemorluto/reaPrepare or rewrite REA release changelogs and GitHub release notes from pinned Git history, with verified contributor thanks and Release Please synchronization. | Stars: 470k | Updated | Claude CodeCodex | Verified | Install |
| academy-guideanthropics/skillsStop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content. | Stars: 197k | Updated | Claude CodeCodex | Verified | Install |
| claude-apianthropics/skillsReference for the Claude API / Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration. TRIGGER — read BEFORE opening the target file; don't skip because it "looks like a one-liner" — whenever: the prompt names Claude/Anthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, `anthropic`, `@anthropic-ai`, `claude-*`, `us.anthropic.*`, `[1m]`); the user asks about an LLM (pricing/model choice/limits/caching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent/MCP/tool-definition/multi-agent/RAG/LLM-judge/computer-use; generate/summarize/extract/classify/rewrite/converse over NL; debugging refusals/cutoffs/streaming/tool-calls/tokens). SKIP only when another provider is being worked on (overrides all triggers): OpenAI/GPT/Gemini/Llama/Mistral/Cohere/Ollama named in the query; OR `grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama'` over the project hits (run this grep FIRST if no provider named — don't Read the file). | Stars: 197k | Updated | Claude CodeCodex | Unverified | Install |
| doc-coauthoringanthropics/skillsGuide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks. | Stars: 197k | Updated | Claude CodeCodex | Verified | Install |