alterlab-tubitak-proposal — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited alterlab-tubitak-proposal (Agent Skill) and scored it 96/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 1 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 1 flagged
A bulleted imperative like {match} tells the agent to never reveal, disclose, or mention something to the user. Used adversarially it can instruct the agent to hide its tool calls or lie about what it did — stripping the transparency a user relies on to trust the agent.
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Scaffolds a TÜBİTAK ARDEB (Araştırma Destek Programları Başkanlığı — the Research Support Programmes directorate) proposal against the verified official form tree, in the directorate's own Turkish section order, then maps the researcher's content into each heading and checks the hard caps the panel rejects on. It owns the Turkish-specific form structure, terminology, and program rules only — generic grant-craft (broader-impacts argumentation, feasibility narrative, Gantt aesthetics) is delegated to alterlab-research-grants.
Two variants:
Programme for Scientific and Technological Research Projects). The heavy, full-form program.
rolling-application variant.
Write me a TÜBİTAK 1001 proposal outline for <topic>
Draft the özgün değer (original value) section for my ARDEB 1001
Scaffold a 1002-A Hızlı Destek başvurusu
Turn my NSF broader-impacts paragraph into a TÜBİTAK yaygın etki section→ Pick the variant, generate the section tree (scripts/scaffold_proposal.py), draft each heading from the user's material, then run the cap/structure check (scripts/scaffold_proposal.py --check) before reporting. Always print the verify-current-call disclaimer — caps and forms change every application period.
Use this skill when the request is to author, outline, or section-map a TÜBİTAK ARDEB 1001 or 1002-A proposal, draft a specific Turkish section (özgün değer = original value; yaygın etki = broader impact/dissemination; yöntem = method; iş paketi = work package), or translate international broader-impacts framing into TÜBİTAK terminology.
The route below sends adjacent requests to the correct sibling skill in the suite. This skill does not check journal indexing, compute career points, or write a non-Turkish grant.
| The request is really about… | Route to |
|---|---|
| Generic grant-craft / NSF / NIH / ERC narrative, broader-impacts argumentation in the abstract | alterlab-research-grants |
| Interim/final progress report (gelişme/sonuç raporu) for an awarded TÜBİTAK project | alterlab-grant-reporting |
| Is a target journal in TR Dizin (national index)? | alterlab-trdizin |
| Does a candidate clear doçentlik (associate-professor) point thresholds? | alterlab-docentlik-eligibility |
| Computing the akademik teşvik (academic-incentive) net score | alterlab-akademik-tesvik |
| The etik kurul (ethics-committee) application / informed-consent form | alterlab-tr-research-ethics |
| The KVKK-compliant data-management plan / TÜBİTAK Veri Yönetim Planı | alterlab-kvkk-dmp |
| Depositing the accepted manuscript on Aperta under the Open Science Policy | alterlab-aperta |
| Turkish APA-7 / TR Dizin manuscript style for the paper itself | alterlab-tr-academic-style |
| Searching DergiPark / YÖK theses for the literature review | alterlab-dergipark, alterlab-yok-tez |
Both forms share the same backbone; 1002-A is the trimmed variant. Section names below are the directorate's own headings (see references/form_structure.md for the full annotated tree and the per-section drafting briefs).
| # | Section (TR) | English gloss | 1001 | 1002-A |
|---|---|---|---|---|
| Özet | ÖZET (TR) + ABSTRACT (EN) | Abstract as two separate blocks (TR + EN), each with Anahtar Kelimeler / Keywords | ✅ | ✅ |
| 1 | ÖZGÜN DEĞER | Original value / significance | ✅ | ✅ |
| 1.1 | Konunun Önemi ve Özgün Değer | Importance & original value of the topic | ✅ | ✅ |
| 1.2 | Araştırma Sorusu / Hipotezi | Research question / hypothesis | ✅ | ✅ |
| 1.3 | Amaç ve Hedefler | Aim & objectives | ✅ | ✅ |
| 2 | YÖNTEM | Method | ✅ | ✅ |
| 3 | PROJE YÖNETİMİ | Project management | ✅ | (light) |
| 3.1 | İş-Zaman Çizelgesi ve İş Paketleri (+ B-Planı) | Work–time chart & work packages (+ contingency plan) | ✅ | ✅ |
| 3.2 | Araştırma Olanakları | Research facilities/resources | ✅ | ✅ |
| 4 | YAYGIN ETKİ | Broader impact / dissemination | ✅ | (light) |
| 4.1 | Öngörülen Çıktılar | Expected outputs | ✅ | ✅ |
| 4.2 | Öngörülen Etkiler / Bilim İletişimi | Expected impacts / science communication | ✅ | ✅ |
| EK-1 | Kaynaklar | References (cited literature) | ✅ | ✅ |
| EK-2 | Bütçe ve Gerekçesi | Budget & its justification | ✅ | ✅ |
| EK-3 | Diğer Projeler / TÜBİTAK Destekleri | Other projects & prior TÜBİTAK support | ✅ | (if applicable) |
Map, don't invent. The directorate evaluates against these headings; never silently restructure them into an IMRaD paper. The özgün değer section is the single most weighted part of a 1001 — it is where the panel decides novelty.
These are administrative-eligibility filters: a proposal that violates them can be returned without scientific review. Pin them with scripts/scaffold_proposal.py --check; see references/program_profiles.md for the full table with as-of dates.
| Item | 1001 | 1002-A |
|---|---|---|
| Özet word cap (TR & EN each) | 600 words | 600 words |
| Project duration | ≤ 36 months | ≤ 12 months |
| Budget upper limit (PTİ/burs hariç) | 3,000,000 TRY (2026-1 period) | 150,000 TRY/yr (incl. burs, as of 2026-02-01) |
| Application window | Periodic call (çağrı) | Rolling / year-round (sürekli) |
| Form length | per current rehber (guide) | ≤ 12 pages excl. annexes |
| PI (Yürütücü) requirement | Doctorate + eligible affiliation | per current rehber |
Verify-current-call disclaimer (always print). Every TRY figure, page limit, and duration above is dated and changes each application period. Before the user submits, WebFetch the live program page and the current başvuru rehberi (application guide) listed in references/program_profiles.md and reconcile. Caps observed for the 2026-1 (1001) and 2026-02-01 (1002-A) periods; TÜBİTAK has announced major 1002 program changes, so treat 1002-A caps as especially volatile.ARDEB panels score against four axes. Draft each section to its axis; the mapping below is the lens the panelist (hakem) uses. Full criteria in references/review_criteria.md.
| Dimension (TR) | English | Carried mainly by |
|---|---|---|
| Bilimsel/Teknolojik Nitelik ve Özgün Değer | Scientific quality & original value | Özet, §1 ÖZGÜN DEĞER, EK-1 |
| Yöntem | Method & feasibility (yapılabilirlik) | §2 YÖNTEM |
| Proje Yönetimi ve Araştırma Olanakları | Project management & resources | §3 (iş paketleri, B-Planı, §3.2) |
| Yaygın Etki | Broader impact & dissemination | §4 YAYGIN ETKİ, EK-2 |
When the user arrives with NSF/NIH/ERC material, translate the concept, not word-for-word. Full glossary in references/terminology_bridge.md.
| International term | TÜBİTAK term | Note |
|---|---|---|
| Significance / innovation / intellectual merit | Özgün değer | The most weighted axis; lead with what is genuinely new |
| Broader impacts | Yaygın etki | Split into çıktılar (outputs), etkiler (impacts), bilim iletişimi (sci-comm) |
| Feasibility | Yapılabilirlik | Lives inside §2 YÖNTEM + §3 management |
| Contingency / risk plan | B-Planı | A required sub-element of the work-package section, not optional |
| Work package / Gantt | İş paketi / İş-Zaman Çizelgesi | Gantt aesthetics → delegate to alterlab-research-grants |
| Aims & objectives | Amaç ve Hedefler | Hedefler should be measurable, tied to work packages |
uv run python skills/turkish-academia/alterlab-tubitak-proposal/scripts/scaffold_proposal.py \
--program 1001 \
--title "<project title>" \
--out proposal_scaffold.md--program accepts 1001 or 1002a. The script emits the full Markdown section tree with the TR heading, its English gloss, the panel dimension each section serves, and a short drafting brief per heading. Use --lang tr (default) or --lang both for bilingual headings.
Fill the scaffold from what the user provides. Keep the directorate's ordering. For özgün değer, state the gap and the new contribution explicitly. For yaygın etki, populate all three sub-parts (çıktılar / etkiler / bilim iletişimi). Pull literature into EK-1 Kaynaklar — and have alterlab-citation-verifier existence-check the bibliography before submission.
uv run python skills/turkish-academia/alterlab-tubitak-proposal/scripts/scaffold_proposal.py \
--check proposal_scaffold.md --program 1001The checker reports: missing required sections, TR/EN özet word counts vs the 600-word cap, whether duration/budget statements (if present) exceed the program ceilings, and whether a B-Planı sub-section exists. It is advisory — it never edits the proposal and it always restates the verify-current-call disclaimer because the ceilings are period-specific.
alterlab-kvkk-dmp,alterlab-aperta); do not draft compliance text here.
ardeb-pbs.tubitak.gov.tr. Proposals are entered andsubmitted here.
arbis.tubitak.gov.tr. A current ARBİS record is aprerequisite; the PI/personnel CVs flow from it. Tell the user to refresh ARBİS before starting the PBS entry.
This skill does not automate or log into either system — it produces the document content the researcher pastes/uploads.
guaranteed cap?
siblings rather than answering them here?
references/form_structure.md — the full annotated 1001 & 1002-A section tree with a draftingbrief per heading and the 1001-vs-1002-A delta.
references/program_profiles.md — per-program caps (duration, budget, window, form length)with as-of dates and the live verification URLs.
references/review_criteria.md — the four panel dimensions and how each section is scored.references/terminology_bridge.md — international ↔ TÜBİTAK concept glossary.Part of the AlterLab Academic Skills suite.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.