alterlab-grant-reporting — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited alterlab-grant-reporting (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
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.
Post-award is where money already won has to be kept: funders release the next tranche only when a compliant progress report lands on time, and they claw back or flag awards when reports slip or numbers don't reconcile. This skill drafts those deliverables — progress/annual reports, final reports, milestone and deliverable tracking, budget-vs-actual narratives, and no-cost-extension (NCE) and rebudgeting justifications — in the exact shape each major funder demands.
It is the post-award counterpart to the pre-award proposal skills: it never writes a new proposal or a fresh budget request. It reports against an award that already exists.
Use this skill when the user needs to:
or Final — section by section.
Project Outcomes Report.
narrative (Part B) and the structured-table / financial-statement framing.
narrative.
by category.
(scripts/report_deadlines.py).
| The request is really about… | Route to |
|---|---|
| Writing a new proposal / project description for NSF, NIH, DOE, DARPA, NSTC (pre-award) | alterlab-research-grants |
| A TÜBİTAK proposal (1001, 1002, etc.) — Turkish funder, pre-award | alterlab-tubitak-proposal |
| Akademik Teşvik annual incentive-points application (ÜAK) | alterlab-akademik-tesvik |
| A data management plan / KVKK-compliant DMP for the proposal stage | alterlab-kvkk-dmp |
| A recommendation / reference letter for a student or colleague | alterlab-recommendation-letters |
| Program / department accreditation assessment & assurance-of-learning reports | alterlab-accreditation-aol |
| Verifying that the citations in the report actually exist | alterlab-citation-verifier |
This skill assumes the award is already made. If the user is still competing for funding, stop and route to the pre-award skill above.
Each funder fixes the structure; do not invent sections. Deep per-funder detail — every section, what goes in it, and which sections apply to which report type — lives in references/funder_formats.md. The essentials:
Three report types share one section skeleton (A–I):
| § | Title | Note |
|---|---|---|
| A | Cover Page | |
| B | Accomplishments | goals, what was done, results, dissemination |
| C | Products | publications, datasets, software, IP |
| D | Participants | personnel + effort; D.1 is the only Section D item required on Interim/Final |
| E | Impact | impact on the field, other disciplines, society |
| F | Changes | changes/problems; rebudgeting & scope changes surface here |
| G | Special Reporting Requirements | |
| H | Budget | Annual (non-SNAP) only — not on Interim/Final |
| I | Outcomes | required on Interim & Final, public-facing |
Report types: Annual (next budget period), Interim (filed with a competing renewal — becomes the Final report if the renewal is not funded), Final (closeout). See the references file before drafting.
cumulative).
800 words**, separate from the final report.
the Portal from the Continuous Reporting module) and Part B (the narrative, mirroring the application form and reporting differences/deviations).
(plus a Certificate on the Financial Statements where the threshold applies).
feeds Part A automatically.
Reporting deadlines are deterministic functions of the award dates, and missing one can suspend payments. Use the helper instead of recalling rules from memory:
uv run python skills/faculty-life/alterlab-grant-reporting/scripts/report_deadlines.py \
--funder nih-snap --budget-period-start 2026-09-01 --project-end 2028-08-31It encodes only the verified rules (sourced in references/funder_formats.md):
non-SNAP ≈ 60 days before; Final RPPR within 120 days of the period-of-performance end.
Project Outcomes Report within 120 days after the award end date.
each reporting period ends.
The script prints the computed due dates as a table or JSON; it does no network I/O and always tells the user to confirm against the funder's own notice of award, because programs can override these defaults.
user hasn't said, ask (NIH vs NSF vs EU; Annual vs Final).
report_deadlines.py from the award dates and leadwith it, so the user knows the time pressure.
what was accomplished this period, products (publications, data, software), personnel and effort, and the budget figures (planned vs actual).
merge or rename funder sections. For each accomplishment, tie it back to a stated aim or deliverable.
planned vs actual amount, the reason, and the corrective action or the rebudgeting/NCE request it motivates. Detail and templates: references/budget_variance.md.
references/nce_rebudget.md: unspent-funds status, the programmatic reason, the revised timeline, and the explicit confirmation that no new scope or new funds are requested.
missing, mark it [TODO: confirm] for the user — do not invent figures, dates, DOIs, or results. If citations are involved, hand the reference list to alterlab-citation-verifier.
publication details, and award numbers come only from the user's materials. Mark anything missing as a [TODO] placeholder.
through eRA Commons / Research.gov / the EU Funding & Tenders Portal. The skill has no portal access and claims none.
general rules; specific programs and cooperative agreements can override them.
final funded year; an NIH Final RPPR adds Section I Outcomes. Don't carry the wrong scope across report types.
references/funder_formats.md — full NIH RPPRA–I section guide (with which sections apply to Annual/Interim/Final), NSF report types, and Horizon Europe / ERC report structure, each with its source.
references/budget_variance.md — budget-vs-actualnarrative template, variance categories, and worked examples.
references/nce_rebudget.md — no-cost-extension andrebudgeting/virement justification structures and the questions each funder asks.
scripts/report_deadlines.py — stdlib-only report-deadline calculator.Part of the AlterLab Academic Skills suite.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.