Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add zongmin-yu/semantic-scholar-skills --skill trace-citationsgit clone --depth 1 https://github.com/zongmin-yu/semantic-scholar-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/zongmin-yu/semantic-scholar-skills/trace-citations)<a href="https://agentmods.dev/skills/zongmin-yu/semantic-scholar-skills/trace-citations"><img src="https://agentmods.dev/badge/skills/zongmin-yu/semantic-scholar-skills/trace-citations/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/zongmin-yu/semantic-scholar-skills/trace-citations"><img src="https://agentmods.dev/badge/skills/zongmin-yu/semantic-scholar-skills/trace-citations.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00027 | $0.00444 |
| Opus 5 | $0.00014 | $0.00222 |
| Sonnet 5 | $0.00005 | $0.00089 |
| Haiku 4.5 | $0.00003 | $0.00044 |
Grade A, and why
trace-citations scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
Trace Citations
Map the citation graph around one focal paper into useful buckets. Use this when the human wants lineage, influence, and strong versus weak citation edges around a paper.
Arguments
- The positional argument is the focal paper query. Quote multi-word titles.
--depth 1|2controls whether to expand a second hop from the strongest first-hop edges.--max-references <n>and--max-citations <n>cap the first-hop fetch sizes.--second-hop-limit <n>caps how many first-hop anchors get expanded at depth two.
Workflow
- Run
python scripts/run.py .... - Read
result.foundationsfor strong references behind the focal paper. - Read
result.direct_descendantsfor strong citing descendants. - Read
result.bridge_nodesfor medium-confidence connectors with rich context or intent signal. - Read
result.weak_edgesfor low-signal edges that are probably less useful. - If
depth=2, readresult.second_hoponly after the first-hop picture looks sensible.
Output
- The script prints the unified JSON envelope described in
output_contract.md. - The underlying workflow result is
CitationTraceResult.to_dict(). result.reference_count_examinedandresult.citation_count_examinedshow the first-hop search breadth.
When To Escalate
- The focal paper resolves incorrectly.
- The API returns very sparse context and intent data, making edge interpretation weak.
- The first-hop graph is too noisy and needs a tighter focal paper choice before going to depth two.
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 44 lines · 27 tokens per session scan A 9759fed2e65b
trace-citations is a skill published in the GitHub repository zongmin-yu/semantic-scholar-skills (21 stars, last pushed 6mo ago), licensed MIT. It adds 27 tokens to every session and 444 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
academic-research
Nested swiss-knife reference for academic literature work — find papers, fetch full-text PDFs, trace citations, write LaTeX manuscripts. First action for any "get me this paper" request: python3 /scripts/fetchpaper.py — walks arXiv → Unpaywall → Europe PMC → CORE → in-house publisher-page extraction…
research-architect
Use when a user needs the end-to-end Research Architect workflow to turn a raw idea, user-designated accessible target reference paper, partial research materials, results, or an existing draft into an auditable research plan, draft, or revision path.
research-architect-draft
Use when a Research Architect project has an evidence-ready spine, evidence and claim trail, citation support, terminology/object identities, and display logic that need to become an evidence-first, promise-aware first draft or audited revision.
research-architect-evidence
Use when executed research materials, results, observations, analyses, or draft claims need to become a Research Architect evidence bank, claim register, manuscript-wide display sequence, panel-level evidence map, and bounded interpretation trail.
research-architect-literature
Use when a Research Architect project needs accessible target-reference papers analyzed for transferable research logic, or candidate questions positioned against literature, gaps, evidence standards, and independent study options—not generic bibliography collection.
research-architect-audit
Use when reviewing, revising, or deciding whether to complete a Research Architect draft, especially when its reference adaptation, design, evidence, citations, claim boundaries, copying risk, or missing upstream artifacts need an auditable diagnosis.