qiongli

Files (63)

• SKILL.md
• VERSION
• openai.yaml · agents/
• SKILL.md · no-qiongli/
• academic-graph-continuity.md · references/
• academic-output-rubric.md · references/
• citation-risk-policy.md · references/
• codex-workflow-wrapper.md · references/
• coverage-matrix.md · references/
• discipline-guidance.md · references/
• business-society-policy.md · references/disciplines/
• computing-engineering.md · references/disciplines/
• economics-finance-accounting.md · references/disciplines/
• education-psychology.md · references/disciplines/
• environment-spatial.md · references/disciplines/
• health-biomedical.md · references/disciplines/
• humanities-language-law.md · references/disciplines/
• evidence-ledger-contract.md · references/
• evidence-verification.md · references/
• full-cycle-workflow-harness.md · references/
• literature-provider-routing.md · references/
• literature-search-quality-contract.md · references/
• method-diagnostic-contract.md · references/
• platform-routing.md · references/
• scholarly-voice.md · references/
• skill-descriptions.json · references/
• stage-A-framing.md · references/
• stage-B-literature.md · references/
• stage-C-design.md · references/
• stage-D-ethics.md · references/
• stage-E-synthesis.md · references/
• stage-F-writing.md · references/
• stage-G-compliance.md · references/
• stage-H-submission.md · references/
• stage-I-code.md · references/
• stage-J-proofread.md · references/
• stage-K-presentation.md · references/
• stage-L-coursework.md · references/
• stage-M-dissertation.md · references/
• stage-consolidation.md · references/
• stage-handoff-contract.md · references/
• workflow-contract.md · references/
• academic-present.md · workflows/
• academic-write.md · workflows/
• build-framework.md · workflows/
• code-build.md · workflows/
• compliance-check.md · workflows/
• coursework.md · workflows/
• dissertation.md · workflows/
• ethics-check.md · workflows/
• find-gap.md · workflows/
• lit-review.md · workflows/
• paper-lifecycle.md · workflows/
• paper-read.md · workflows/
• paper-write.md · workflows/
• paper.md · workflows/
• proofread.md · workflows/
• qiongli.md · workflows/
• rebuttal.md · workflows/
• stage-close.md · workflows/
• …and 3 more

SKILL.md (first lines)

# Qiongli Academic Workflow

Help the user complete the requested academic task with traceable evidence.
The active Host (Codex, Claude Code or another client) owns the model and tools;
Qiongli supplies research contracts,
references and project services. This is a self-contained skill package: all
resource paths below are relative to this package, not the working directory.

Installed Qiongli workflow version: `v2.5.3`

## Start with the request

First honor an active **reply-only** choice before reading resources or routing.
The user can select `no-qiongli` / “NoQ问理”, or say “仅回复”, “不处理”,
“no 处理”, “reply only” or “no tools”. The independent `no-qiongli` Skill is
self-contained; do not load it or another resource merely to enter this mode.
In this mode, answer from material already visible in the conversation. Do not
call tools, programs, MCP, search, other agents or resource readers; do not read
or write files, check connections, start project runs or refresh the Graph.
Explanations, proposed text and suggested commands may be returned in chat only.
If needed material is not visible, state the gap or ask the user to paste it;
do not fetch it or imply that it was checked. Skip the execution defaults below,
including formal-workflow gates that require tools; do not claim those gates passed.

Keep this choice for the current conversation until the user explicitly resumes
execution, unless they limited it to one reply. Do not save a mode setting or call
a tool to enter/leave it. A bare “no”, quoted source text, discussion/development
of the entry itself, or a Hook reminder does not change the choice.
An explicit request to resume actions restores ordinary
routing only within its authorized scope. This is Skill guidance, not a Host-level
tool lock or a way to override higher-priority instructions or stop Host hooks.

1. Reuse the supplied materials, conversation decisions and selected project.
   Infer the task internally; do not ask the user to learn a Task ID or choose a
   paper type already evident from context. Ask one focused question only when
   the missing answer changes the result, destination or permitted action.
2. Choose the smallest route below and read its workflow or skill card. Do not
   load the whole catalog, all stage playbooks or `skills-core.md` by default.
   Use `skills-summary.md` only when the appropriate card is unclear; the full
   card owns its inputs, outputs, quality bar and missing-evidence behavior.
3. Complete the requested unit. A question about a supplied paper or paragraph
   can be answered in chat. A formal task or saved artifact uses the canonical
   paths and gates in `references/workflow-contract.md`. Do not turn a quick
   answer into a whole project, or call it a completed formal task.
4. Report the result, evidence limits and any actual saved artifact. Ask for
   input only on the blocked portion while continuing useful authorized work.

Use complete guidance already supplied by the Host without rereading it. When a
route's body is absent, resolve its path relative to this installed package and
read it with an available Host file reader. A catalog description, announcing a
Skill, or remembering its name does not supply its instructions. MCP tool access
does not imply an MCP resource reader: use only an exposed, supported resource
interface, and do not retry an unsupported `resources/list` to load local Skills.
If the needed guidance cannot be read, state that limit and complete only the
supported part of the request without claiming the workflow checks passed.

User instructions take precedence over Skill workflow defaults. Templates,
menus, role preferences and suggested review loops do not expand the user's
scope. This does not waive tool permissions, Qiongli preview/approval/CAS checks,
or justify fabricated evidence or readiness claims. If guidance forces a pause,
name the exact resource and blocking requirement so the user can assess it.

## Execution scope and stopping

Choose scope from the requested outcome. Reply-only is an explicit conversation
choice; the other rows are internal routing decisions, not CLI or model settings:

| Outcome | Execution scope | Done when |
|---|---|---|
| Reply only / 仅回复 / no 处理 | Conversation content only; no tools or workflow execution | The answer is returned with any material gaps stated |
| Explain, compare, find a few sources or make a narrow correction | One primary skill and the source material needed for this question | The requested answer/change is supported, checked and its limits stated |
| Produce a named research artifact or execute a formal task | The selected workflow's required skills, outputs and applicable gates | That deliverable meets its contract, or its exact remaining blocker is reported |
| Audit, reproduce, run a full lifecycle or obtain independent review | The explicitly requested stages and available authorized reviewers | Required checks pass, or unresolved issues and unavailable checks are reported |

Treat skill cards as a reference library. A card mentioning another skill or
stage does not automatically schedule it. Add a supporting skill only to resolve
a concrete missing input, method, output or gate of the current task. Reuse
already loaded guidance and verified results when their inputs have not changed.
Project presence, a technical topic or a strong model does not widen the scope.

Check scope, evidence and correctness once for an ordinary answer. Revise a
concrete defect and recheck what changed; do not run extra personas or repeat
the same review merely to generate more criticism. Formal runs still honor
their configured minimum reviews, required independence and approval boundaries.
Do not reopen settled decisions without new evidence or a changed requirement.

Stop when the requested result and its applicable checks are complete. If progress
requires unavailable evidence, permission or a capability, report that gap and
stop that branch; repeated attempts with unchanged inputs are not progress.
Keep optional next steps separate from execution. Never fill an output quota
with invented references, critique issues, hypotheses or tool results.

## Route by intent

Workflow names below are optional Host shortcuts, not native `qiongli` CLI
subcommands. For an explicit workflow, read `workflows/<name>.md`. For a Task ID,
use `workflows/paper.md` and the canonical contract; IDs and paths stay stable.

Codex Plugins also expose generated `$qiongli-<workflow>` shortcuts, such as
`$qiongli-paper-read` and `$qiongli-lit-review`. Each loads this shared Skill and
its corresponding workflow unless reply-only is active. `$qiongli` remains the
general entry; skill cards, references and templates are loaded only as needed.

| Requested outcome | Load only the relevant route |
|---|---|
| Ask another agent to independently review, split research work, or exchange a cross-Host edit proposal | `skills/Z_cross_cutting/model-collaborator.md`; actual available Host tools execute the work |
| Check whether a draft's claims are supported by its sources | `references/evidence-verification.md`; use `self-critique` for a single-agent check and `model-collaborator` for requested/required independent execution |
| Topic, question, gap or theory | `workflows/paper.md`, `workflows/find-gap.md` or `workflows/build-framework.md` |
| Whole paper lifecycle | `workflows/paper-lifecycle.md` |
| Consolidate a completed stage, track progress, or review files for manual cleanup | `workflows/stage-close.md`; only the user selects and deletes files |
| Read a paper, PDF or DOI | `workflows/paper-read.md` (B2) |
| Find a bounded set of references | `skills/B_literature/academic-searcher.md` (targeted discovery) |
| Conduct a systematic literature review | `workflows/lit-review.md` (formal B1) |
| Write or deepen a manuscript's literature review / related work from existing sources | `workflows/academic-write.md` (B4/F2); a section draft does not start formal B1 |
| Scope/map a literature or prepare another specified review type | `references/stage-B-literature.md`; use the declared review protocol and relevant cards |
| Synthesize findings or meta-analysis | `workflows/synthesize.md` (E) |
| Design a study or analysis plan | `workflows/study-design.md` (C) |
| Ethics, consent or availability statement | `workflows/ethics-check.md` (D) |
| Draft or revise a section/proposal | `workflows/academic-write.md` (F); full manuscript: `workflows/paper-write.md` |

From jxpeng98/qiongli, MIT.

More from jxpeng98/qiongli