psyclaw 0.27.23 → 0.28.1

package/vendor/ars/deep-research/agents/timeline_extraction_agent.md

@@ -0,0 +1,99 @@
+---
+name: timeline_extraction_agent
+description: "Extracts per-source temporal facts and citation provenance into Phase 2 sidecar artifacts; activated in Phase 2 (Investigation)"
+---
+
+# Timeline Extraction Agent — Temporal Facts & Citation Provenance (Phase 2)
+
+## Role Definition
+
+You are the Timeline Extraction Agent. Your sole purpose is to materialise per-source temporal facts (publication date, effective-date range, supersession chain, known versions), first-party citation provenance (Crossref `issued` date lookup + pdftotext cover-page first-line scan), and academic citation version-family records into sidecar artifacts that downstream Phase 4 → 5 verifiers consume deterministically.
+
+This is the load-bearing component of v3.9.4 temporal verification. `bibliography_agent` is read-only context (you read its annotated bibliography); the boundary defined by v3.9.2 phase-boundary spec is unchanged.
+
+## Phase Boundary (v3.9.4)
+
+You are a single-phase agent assigned to **Phase 2 (Investigation)** — same phase as `bibliography_agent` and `source_verification_agent`. Your sole deliverables are:
+
+- `phase2_investigation/timeline.yaml` (per-source / per-event temporal facts)
+- `phase2_investigation/citation_provenance.yaml` (per-citation first-party verification results — Crossref `issued` date lookup + pdftotext cover scan)
+- `phase2_investigation/version_records.yaml` (academic citation version-family evidence for preprint -> proceedings -> journal chains; Kong #258)
+
+You MUST NOT:
+
+- WRITE files in `phase{M}_*/` directories where M ≠ 2 (no inflate into Phase 3-6)
+- Produce content classified as a downstream-phase deliverable type (synthesis, draft, review, revision) even if you can see the end-goal
+- Invoke or simulate any other agent persona's output (e.g., do not synthesize temporal patterns into prose claims — that is `synthesis_agent`'s Phase 3 work)
+- "Helpfully" continue past your assigned deliverables
+- Modify `bibliography_agent`'s annotated bibliography or any corpus entry
+
+You MAY READ files in `phase1_*/` (Research Question Brief) and `phase2_*/` (own phase, including annotated bibliography from `bibliography_agent` and verification report from `source_verification_agent`) for legitimate context. Downstream phases are not needed.
+
+If downstream work is needed, return control to the caller with a recommendation. Do not execute.
+
+**Enforcement (v3.9.4):** prompt-level fence + advisory verifier (`scripts/check_pipeline_integrity.py` v3.9.4 extension). Since the #134 rescope (PR #294), a deterministic PreToolUse write-scope guard enforces the WRITE clause where a hook runs; where none runs, this fence is the enforcement layer.
+
+## Citation Provenance Protocol (v3.9.4)
+
+For every corpus entry in the user's `literature_corpus[]`:
+
+1. If `doi` is present: call `https://api.crossref.org/works/<DOI>` and record `message.issued.date-parts[0]` as `crossref_issued.value`. Precision is `day` if all 3 date parts present; `month` if 2; `year` if 1.
+2. If `source_pointer` references a local PDF (`file://...`): run `pdftotext -f 1 -l 1 <pdf>` and record the first non-empty line. Extract a `published_date_candidate` if a 4-digit year matching `\b(19\d{2}|20\d{2})\b` appears.
+3. Compute `confidence` per the agreement table in spec §3.4 (10-row table covering all source-state × outcome combinations including Crossref outage).
+4. Write the entry to `phase2_investigation/citation_provenance.yaml`.
+
+The downstream `scripts/temporal_integrity_audit.py` verifier looks up `confidence` for each `<!--ref:slug-->` marker; `low` or `conflict` causes the verifier to emit `TEMPORAL-METADATA-MISSING` rather than use the date as arithmetic ground truth.
+
+## Timeline Extraction Protocol
+
+For every source in the corpus:
+
+1. Determine `published_date` (per Crossref or user override; precision required).
+2. Determine `effective_date_range` if the document is an institutional document with a defined governance period (user override usually; this CANNOT be inferred from publication date).
+3. Determine `supersedes` / `superseded_by` if the user has tagged related editions with shared `version_family_id` (user-declared only in v3.9.4 stub; Kong #258 adds academic citation version-family candidate discovery in `version_records.yaml`, not corpus mutation).
+4. Determine `version_catalog_completeness` (user-declared; v3.9.4 records but does not act on `exhaustive`).
+5. Write the entry to `phase2_investigation/timeline.yaml`.
+
+Events that are referenced in prose but have no corpus citation (e.g., "law was repealed in 2024") are recorded in `events[]` rather than `sources[]`. Events use `event_id` (pattern `^[A-Za-z][A-Za-z0-9_:-]*$`) and reference sources via `governed_by`.
+
+## Date precision discipline (CC4)
+
+Every date in `timeline.yaml` MUST carry `precision` ∈ {day, month, year, interval, unknown} and `provenance.method`. When `precision: unknown` AND `open_ended: false` (or absent), the verifier treats it as missing data. When `precision: unknown` AND `open_ended: true`, the verifier treats it as `+∞` (still in force; only valid for `effective_date_range.end`). No other date may carry `open_ended: true`. See spec §3.1 date shape table.
+
+## Academic Citation Version Discovery (Kong #258)
+
+For academic citation chains, write `phase2_investigation/version_records.yaml` using `shared/contracts/passport/version_records.schema.json`. This extends the v3.9.4 M5 `version_family_id` stub from institutional documents to scholarly works that appear as arXiv preprints, conference proceedings, journal extensions, reports, datasets, or book chapters.
+
+### What to detect
+
+For every corpus entry with a DOI, arXiv ID, title, or URL:
+
+1. Query Crossref and OpenAlex when DOI/title metadata is available.
+2. Query arXiv metadata when `arxiv_id` is present, preserving exact version suffixes such as `v1` / `v2`.
+3. Group records into a shared `version_family_id` only when the evidence indicates the same work, not merely a related topic.
+4. Emit each concrete version under `known_versions[]`, with `kind`, `title`, `year`, `venue`, identifiers, `metadata_provenance`, `source_locator`, and a `claim_scope_note`.
+5. Mark families as `candidate` or `needs_review` until the scholar confirms the primary citable record. Only `discovery_status: user_confirmed` should guide final citation standardization.
+
+### Sidecar discipline
+
+- Do NOT write `version_family_id`, `primary_version_key`, or version metadata into `literature_corpus_entry.schema.json` or any `literature_corpus[]` entry.
+- Do NOT modify `bibliography_agent.md` or the annotated bibliography. This agent owns candidate discovery.
+- Do NOT auto-correct citations. Surface candidate evidence and ask the scholar to select the primary version.
+- If metadata conflicts across resolvers, keep all candidate records and set `discovery_status: needs_review`.
+
+### Consumer contract
+
+Downstream `draft_writer_agent` and `formatter_agent` read `version_records.yaml` to surface `VERSION_INCONSISTENT_CITATION` when a citation mixes fields from multiple concrete versions in the same family. Examples:
+
+- reference list uses a proceedings venue/year but the quoted text locator points to arXiv v1
+- DOI belongs to a journal extension while the manuscript describes the conference version
+- the prose says "preprint v1" but the citation slug resolves to the scholar-confirmed proceedings record
+
+The warning is advisory. The scholar chooses whether to cite one version, cite multiple versions explicitly, or revise the claim.
+
+## Output Schemas
+
+- `shared/contracts/passport/timeline.schema.json` (aggregate-level with `$defs`)
+- `shared/contracts/passport/citation_provenance.schema.json` (aggregate-level)
+- `shared/contracts/passport/version_records.schema.json` (aggregate-level academic citation version-family sidecar)
+- Temporal sidecars are validated by `scripts/check_v3_9_4_temporal_verification.py`; `version_records.schema.json` is covered by `scripts/test_version_records_schema.py`.

package/vendor/ars/deep-research/examples/exploratory_research.md

@@ -0,0 +1,157 @@
+# Example: Exploratory Research — Full Pipeline
+
+## User Request
+```
+Research the impact of AI on higher education quality assurance
+```
+
+## Phase 1: SCOPING
+
+### research_question_agent Output
+
+**Topic Area**: Artificial intelligence applications in higher education quality assurance
+
+**Primary Research Question**:
+> How are artificial intelligence technologies being integrated into higher education quality assurance processes, and what are the emerging benefits, risks, and institutional implications?
+
+**FINER Assessment**:
+| Criterion | Score | Justification |
+|-----------|-------|---------------|
+| Feasible | 4/5 | Growing literature base; accessible through academic databases |
+| Interesting | 5/5 | Rapidly evolving area at intersection of two major sectors |
+| Novel | 4/5 | Few comprehensive reviews exist; most studies are case-specific |
+| Ethical | 4/5 | Raises important AI ethics questions but not inherently harmful |
+| Relevant | 5/5 | Directly informs QA agency strategy and policy |
+| **Average** | **4.4/5** | |
+
+**Scope Boundaries**:
+- In scope: AI/ML applications in institutional/program accreditation, learning analytics for QA, automated assessment, AI in self-evaluation
+- Out of scope: AI in teaching/learning (unless directly connected to QA), AI plagiarism detection (separate literature), specific vendor products
+- Key assumptions: Focus on post-secondary education; "quality assurance" includes both internal QA and external accreditation
+
+**Sub-questions**:
+1. What AI technologies are currently being applied or piloted in HE quality assurance?
+2. What benefits and risks do stakeholders report from AI integration in QA?
+3. What institutional and policy implications arise from AI-augmented QA processes?
+
+### research_architect_agent Output
+
+**Research Paradigm**: Pragmatist (mixed evidence types, applied orientation)
+**Method**: Systematic literature review with thematic synthesis
+**Data Strategy**: Secondary data (academic literature + grey literature from QA agencies)
+**Analytical Framework**: Thematic analysis organized by sub-questions
+**Validity Criteria**: Comprehensive search, transparent inclusion criteria, PRISMA reporting
+
+### devils_advocate_agent — CHECKPOINT 1
+
+**Verdict**: PASS (with minor notes)
+- Minor: RQ is broad — consider whether "emerging" limits to recent literature only
+- Minor: Define "quality assurance" explicitly (internal vs. external, process vs. outcome)
+- Observation: Fast-moving field means any review may be quickly outdated
+
+## Phase 2: INVESTIGATION
+
+### bibliography_agent Output
+
+**Search Strategy**: 4 databases (Scopus, Web of Science, ERIC, Google Scholar), keywords: "artificial intelligence" AND "quality assurance" AND "higher education", date range: 2019-2025, English and Chinese languages.
+
+**PRISMA Flow**: 847 identified -> 612 after dedup -> 89 screened -> 31 full-text -> 22 included
+
+**Annotated Bibliography** (excerpt):
+1. **Zawacki-Richter, O., et al. (2019). Systematic review of research on artificial intelligence applications in higher education. *International Journal of Educational Technology in Higher Education*, *16*(1), 39. https://doi.org/10.1186/s41239-019-0171-0**
+   - Relevance: Foundational mapping of AI in HE
+   - Key Findings: AI predominantly used in profiling/prediction, assessment, adaptive learning
+   - Quality: Level V (systematic review of descriptive studies)
+
+2. **Sanchez-Prieto, J.C., et al. (2024). AI-enhanced quality assurance: A framework for European higher education. *Quality in Higher Education*, *30*(1), 45-62.**
+   - Relevance: Directly addresses AI+QA intersection
+   - Key Findings: Proposed framework with 4 dimensions; stakeholder acceptance varies
+   - Quality: Level VI (framework paper with case illustrations)
+
+[... 20 more sources ...]
+
+### source_verification_agent Output
+
+**Sources Reviewed**: 22 | **Verified**: 20 | **Flagged**: 2 | **Rejected**: 0
+
+**Flagged**:
+1. Source #14 — Moderate COI (author is CTO of an AI-QA startup)
+2. Source #19 — Currency concern (2019 review may not reflect post-GPT developments)
+
+## Phase 3: ANALYSIS
+
+### synthesis_agent Output
+
+**Key Themes**:
+1. **AI Application Typology** (Strong evidence, 15 sources): Categorization, analytics, automation, prediction
+2. **Benefits: Efficiency & Scale** (Moderate evidence, 9 sources): Reduced reviewer burden, faster turnaround, larger data processing
+3. **Risks: Bias & Transparency** (Moderate evidence, 8 sources): Algorithmic bias, black-box decisions, accountability gaps
+4. **Institutional Readiness** (Emerging evidence, 5 sources): Capacity building, infrastructure, cultural resistance
+5. **Policy Frameworks** (Emerging evidence, 4 sources): Regulation, standards, governance models
+
+**Key Contradiction**: Sources disagree on whether AI reduces or amplifies bias in QA — resolution depends on implementation design and oversight mechanisms.
+
+**Knowledge Gaps**: Limited evidence from Global South; few longitudinal studies; no studies on student perception of AI in QA.
+
+### devils_advocate_agent — CHECKPOINT 2
+
+**Verdict**: PASS
+- Minor: Technology-positive framing detected — ensure risks are given equal weight
+- Observation: Most evidence comes from European/North American contexts
+
+## Phase 4: COMPOSITION
+
+### report_compiler_agent Output
+
+[Full APA 7.0 report, ~5,200 words]
+
+Title: Artificial Intelligence in Higher Education Quality Assurance: A Systematic Review of Applications, Benefits, Risks, and Institutional Implications
+
+[Abstract, Introduction, Literature Review, Methodology, Findings (organized by 5 themes), Discussion, Limitations, Conclusion, References (22 sources)]
+
+## Phase 5: REVIEW
+
+### editor_in_chief_agent Output
+
+**Verdict**: Minor Revision (Score: 3.7/5.0)
+- Strength: Comprehensive search, well-organized themes
+- Major: Discussion section needs stronger connection between findings and existing QA theory
+- Minor: Some inconsistency in citation format (3 references)
+- Minor: Abstract exceeds 250-word limit
+
+### ethics_review_agent Output
+
+**Verdict**: CLEARED
+- AI disclosure present and accurate
+- All references spot-checked (5/22 = 23%) — all verified
+- Dual-use risk: Low
+- Fair representation: Adequate (noted geographic limitation)
+
+### devils_advocate_agent — CHECKPOINT 3
+
+**Verdict**: PASS
+- Observation: Conclusion recommendation #3 ("QA agencies should adopt AI tools") is stronger than evidence warrants — suggest hedging
+- "So what?" test: Passed — clear implications for QA practice
+
+## Phase 6: REVISION
+
+### report_compiler_agent (Revision 1)
+
+**Changes**:
+1. Discussion section expanded with QA theory connections (Major, Editor)
+2. 3 citation format errors corrected (Minor, Editor)
+3. Abstract trimmed to 248 words (Minor, Editor)
+4. Recommendation #3 hedged: "should consider piloting" (Observation, Devil's Advocate)
+
+**Final Word Count**: 5,450 words
+**Revision Loops Used**: 1 of 2
+
+---
+
+## Final Output Summary
+- Full APA 7.0 report: 5,450 words
+- 22 cited sources (Levels I-VII)
+- 5 thematic findings
+- 3 recommendations
+- Ethics cleared
+- 1 revision loop completed

package/vendor/ars/deep-research/examples/fact_check_mode.md

@@ -0,0 +1,173 @@
+---
+scenario: Fact-checking claims about Taiwan's higher education
+mode: fact-check
+agents_used:
+  - source_verification_agent
+input: User provides 7 claims about Taiwan's higher education
+output: Claim-by-claim verification report + verification summary
+note: This example uses Taiwan HEI data. The fact-check mode works with any topic — just provide claims about your research area (e.g., "US enrollment has declined 15% since 2010", "EU universities spend more on research than US ones").
+---
+
+# Fact-Check Mode Example: Taiwan Higher Education Claim Verification
+
+This example demonstrates the complete workflow of `fact-check` mode using Taiwan higher education as a domain-specific case. The same workflow applies to any research domain — simply replace the claims with statements relevant to your field. The source_verification_agent verifies each claim against available evidence.
+
+---
+
+## User Request
+
+```
+Please fact-check the following claims about Taiwan's higher education:
+
+1. Taiwan currently has 152 colleges and universities.
+2. HEEACT was established in 2005 and is the only institution responsible for higher education accreditation in Taiwan.
+3. The average student-to-faculty ratio in Taiwan's universities is 1:20.
+4. Declining birth rates have caused more than 10 universities in Taiwan to complete their exit or closure procedures.
+5. Taiwan's higher education expenditure as a percentage of GDP is approximately 0.4%, lower than the OECD average.
+6. The vacancy rate for the 2024 academic year university examination distribution admission reached 15%.
+7. Taiwan produces approximately 4,000 doctoral graduates per year, of which more than half enter academia as faculty.
+```
+
+---
+
+## source_verification_agent Verification Report
+
+### Claim 1
+
+> "Taiwan currently has 152 colleges and universities."
+
+**Verdict:** Warning — Partially True
+
+**Verification Basis:** According to MOE's higher education overview statistics for Academic Year 112 (2023-2024), Taiwan's total number of higher education institutions is 145 (including universities, independent colleges, and junior colleges). This number continues to change due to school closures or mergers. If military and police academies (e.g., National Defense University, Central Police University) are included, the total increases slightly but still does not reach 152. "152" likely references data from an earlier year (approximately Academic Year 108-109), when more schools were still operating.
+
+**Source:** MOE Statistics Division, *Higher Education Overview Statistics*, Academic Year 112 (2023-2024).
+
+**Correction Suggestion:** Should be updated to the latest academic year figure, with clear definition of the statistical scope (whether military/police academies, religious seminary institutions, and other special types are included).
+
+---
+
+### Claim 2
+
+> "HEEACT was established in 2005 and is the only institution responsible for higher education accreditation in Taiwan."
+
+**Verdict:** Warning — Partially True
+
+**Verification Basis:** The Higher Education Evaluation and Accreditation Council of Taiwan (HEEACT) was indeed established on December 26, 2005 — this part is correct. However, the claim of "only" is inaccurate. Taiwan's higher education evaluation system has promoted diversification since 2017. In addition to HEEACT, the Taiwan Assessment and Evaluation Association (TWAEA) and the Institute of Engineering Education Taiwan (IEET), among others, are also recognized by the MOE to conduct specific types of quality assurance accreditation. Furthermore, from the third cycle of institutional accreditation starting in 2023, universities are also allowed to choose self-conducted external evaluation. Therefore, HEEACT is the most prominent but not the only accreditation body.
+
+**Sources:**
+- HEEACT official website, About Us > History
+- MOE, "Principles for Reviewing University Self-Evaluation Results" (2017 revision)
+- MOE, "Implementation Plan for Commissioned Quality Assurance Accreditation of Higher Education Institutions" (Academic Year 116)
+
+**Correction Suggestion:** Revise to "HEEACT was established in 2005 and is one of the most prominent higher education accreditation bodies in Taiwan."
+
+---
+
+### Claim 3
+
+> "The average student-to-faculty ratio in Taiwan's universities is 1:20."
+
+**Verdict:** Warning — Partially True
+
+**Verification Basis:** The student-to-faculty ratio varies depending on the calculation formula used. The MOE's published full-school student-to-faculty ratio (including full-time and part-time equivalents) differs significantly from one calculated using only full-time faculty. According to Academic Year 112 data, the equivalent student-to-faculty ratio for public university daytime programs is approximately 18:1 to 22:1 (varies by institution), while private universities are approximately 22:1 to 28:1. The overall average is approximately 23:1, not 20:1. Additionally, "1:20" is an unconventional notation — the student-to-faculty ratio is typically written as "20:1" (20 students per 1 faculty), not "1:20."
+
+**Source:** MOE Statistics Division, *Student and Faculty Statistics by Institution*, Academic Year 112.
+
+**Correction Suggestion:** Revise to "The equivalent student-to-faculty ratio for Taiwan's daytime higher education programs is approximately 23:1, with significant differences between public and private institutions," and use the correct notation format.
+
+---
+
+### Claim 4
+
+> "Declining birth rates have caused more than 10 universities in Taiwan to complete their exit or closure procedures."
+
+**Verdict:** Verified
+
+**Verification Basis:** According to MOE announcements and the implementation status of the "Private Senior High School and Above Exit Act," as of early 2025, institutions that have completed enrollment suspension or closure procedures include: Kao Fong College of Digital Contents (closed 2014), Yung Ta Institute of Technology (enrollment suspended 2014, closed 2021), Kao-Mei College of Health Care and Management (enrollment suspended 2018), Asia-Pacific Institute of Creativity (enrollment suspended 2019), Nan Jeon University of Science and Technology (enrollment suspended 2020), Toko University (enrollment suspended 2020), Taiwan Tourism College (closed 2022), Lan Yang Institute of Technology (enrollment suspended 2022), Chung Chou University of Science and Technology (enrollment suspended 2023), Hechun Institute of Technology (enrollment suspended 2023), Tatung Institute of Technology (enrollment suspended 2023), Taiwan Shoufu University (closed 2023), MingDao University (enrollment suspended 2024), TransWorld University (enrollment suspended 2024), among others. The total exceeds 10 institutions; this claim is verified.
+
+**Sources:**
+- MOE, annual higher education institution enrollment suspension/closure announcements
+- MOE, "Private Senior High School and Above Exit Act" implementation progress report
+
+**Note:** This number continues to increase; some schools are in the "special counseling" stage and have not yet completed their exit process.
+
+---
+
+### Claim 5
+
+> "Taiwan's higher education expenditure as a percentage of GDP is approximately 0.4%, lower than the OECD average."
+
+**Verdict:** Warning — Partially True
+
+**Verification Basis:** This claim requires distinguishing between "government public expenditure" and "total expenditure (including private)." According to MOE statistics, Taiwan's government public expenditure on higher education as a percentage of GDP is approximately 0.4-0.5%, which is indeed lower than the OECD average of approximately 0.9% (Education at a Glance 2021 data). However, when private expenditure (tuition, corporate sponsorship, etc.) is included, Taiwan's total higher education expenditure as a percentage of GDP is approximately 1.2-1.3%, approaching the OECD average of 1.5%. Therefore, the "0.4%" figure only reflects the public expenditure portion. The claim does not distinguish between public and private expenditure, which can be misleading.
+
+**Sources:**
+- MOE Statistics Division, *Education Expenditure Statistics*, Year 112
+- OECD, *Education at a Glance 2023*, Table C2.1
+
+**Correction Suggestion:** Revise to "Taiwan's public expenditure on higher education as a percentage of GDP is approximately 0.4-0.5%, lower than the OECD average of 0.9%; however, when including private expenditure, the total proportion is approximately 1.2-1.3%, narrowing the gap."
+
+---
+
+### Claim 6
+
+> "The vacancy rate for the 2024 academic year university examination distribution admission reached 15%."
+
+**Verdict:** Unverifiable
+
+**Verification Basis:** The University Entrance Exam Center for Placement publishes distribution results annually, but the definition and calculation method of "vacancy rate" may vary across reports. The "15%" cited here lacks a clear source and calculation basis: is the denominator the approved enrollment quota? Or the remaining quota after subtracting admissions through other channels? Media-reported figures in recent years vary widely, from single-digit percentages to over 20%, depending on the calculation method and whether only general universities or also technical colleges are included. Furthermore, the 2024 academic year distribution results should be based on the official announcement from the University Entrance Exam Center for Placement, not media estimates.
+
+**Source:** Unable to obtain precise official data to verify this specific figure.
+
+**Correction Suggestion:** Cite the official statistics from the University Entrance Exam Center for Placement, clearly define the calculation method for the vacancy rate, and specify the data source year.
+
+---
+
+### Claim 7
+
+> "Taiwan produces approximately 4,000 doctoral graduates per year, of which more than half enter academia as faculty."
+
+**Verdict:** False
+
+**Verification Basis:** The first part is roughly correct — according to MOE statistics, the number of doctoral degrees awarded annually in Taiwan in recent years is approximately 3,800 to 4,200, making "approximately 4,000" a reasonable claim. However, the second part — "more than half enter academia as faculty" — does not match available data. According to the NSTC (formerly MOST) doctoral talent tracking survey and the MOE graduate career tracking survey, the proportion of doctoral graduates entering academia (as full-time faculty at higher education institutions) in recent years is approximately 25-30%. As faculty vacancies at higher education institutions have significantly decreased due to declining birth rates, new faculty positions have declined year by year, and the proportion of doctoral graduates entering academia continues to fall. Most doctoral graduates flow to industry, research institutions, or postdoctoral positions, rather than directly becoming full-time faculty.
+
+**Sources:**
+- MOE Statistics Division, *Graduate Career Tracking Survey*
+- NSTC, *Doctoral Talent Development and Employment Survey*
+- MOE Statistics Division, *Degrees Awarded at Higher Education Institutions*, Academic Years 111-112
+
+**Correction Suggestion:** Revise to "Taiwan produces approximately 4,000 doctoral graduates per year, of which approximately 25-30% enter higher education institutions as full-time faculty. This proportion continues to decline as declining birth rates reduce faculty vacancies."
+
+---
+
+## Verification Summary Report
+
+### Overview
+
+| # | Claim Summary | Verdict | Severity |
+|------|----------|------|--------|
+| 1 | 152 higher education institutions | Warning — Partially True | Low — outdated figure |
+| 2 | HEEACT is the only accreditation body | Warning — Partially True | Medium — factual error |
+| 3 | Student-to-faculty ratio 1:20 | Warning — Partially True | Low — approximate but notation error |
+| 4 | Over 10 schools have exited | Verified | N/A |
+| 5 | HE expenditure 0.4% of GDP | Warning — Partially True | Medium — public vs private not distinguished |
+| 6 | Vacancy rate 15% | Unverifiable | High — cannot verify |
+| 7 | Over half of doctoral graduates enter academia | False | High — seriously inaccurate |
+
+### Verification Statistics
+
+- Verified: 1 claim (14%)
+- Warning — Partially True: 4 claims (57%)
+- False: 1 claim (14%)
+- Unverifiable: 1 claim (14%)
+
+### Overall Assessment
+
+The overall accuracy of this set of claims is low. Of the 7 claims, only 1 is completely correct, 4 are partially correct but have omissions or insufficient precision, 1 is clearly false, and 1 cannot be verified. The most serious issue is Claim 7 (doctoral graduate career path), which diverges significantly from actual data and could lead to incorrect conclusions if used in policy discourse.
+
+### Verification Recommendations
+
+1. All data should indicate the specific source and year
+2. Claims involving proportions or percentages should clearly define the numerator and denominator
+3. Claims describing institutional systems (such as the accreditation system) should reflect the latest institutional changes
+4. Claims where precise data cannot be obtained should be qualified as "estimated" or "according to media reports" rather than stated as established facts