- Home
- Skills
- Research & Analysis
- Callimachus — Citation Verification
Works with the AI Agents you already use
Callimachus — Citation Verification
Check manuscript citations from your AI agent: verify sources, check citation style, and assess claim support with traceable source passages.
Free
About this skill
Check citations from your AI agent
Callimachus lets your agent run citation verification from the command line and return a report that connects its findings to retrieved source passages.
What it checks
- Sources: whether each reference exists and matches the cited bibliographic details.
- Citation style: whether references follow the detected or selected style.
- Claim support: whether retrieved source text supports the claims that cite it, with quotations grounded in that text.
How it works
Give your agent a manuscript. The skill guides it through setup, provider configuration, and the full Callimachus workflow: Parse → Resolve → Fetch → Verify → Report.
Choose the accuracy level and model provider. Callimachus combines deterministic validation and passage grounding with an LLM jury for claim relevance. Your agent can help retrieve missing source material through the runtime's task workflow.
What you get
A structured report with source checks, citation-style findings, claim-support verdicts, and the passages behind them. Retrieval gaps and unresolved sources remain visible, so you can focus your review on the citations that need attention.
Requirements
The skill guides setup of the full public Callimachus runtime, from a source archive or a repository clone. Git is optional. Source execution uses CPython 3.13 and network access. Supported CLI backends can reuse your existing Claude, Codex or Gemini CLI access without a separate LLM API key in Callimachus. Source-service API keys are optional; direct API backends use the credentials required by that provider.
Audit-ready mode requires a separately deployed integrity authority. The skill package alone does not provide it.
- The skill package is text-only and does not contain the Callimachus engine, Python dependencies, native libraries, provider credentials, or an integrity authority.
- Source retrieval depends on network access, provider availability, access rights, and rate limits; some sources may remain unavailable or unresolved.
- Claim relevance is assessed by an LLM. Exact passage grounding and deterministic gates reduce fabrication risk but do not guarantee semantic correctness.
- An ordinary completed report is not automatically audit-ready; only a final projection explicitly marked audit_ready=true supports that description.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 14 days ago
- Free with an Agensi account
Requires
Permissions
Uses the user-selected manuscript and its bibliography, a dedicated Callimachus runtime and virtual environment, and generated run/report files. Source retrieval and model-provider requests require network access. Credentials stay in the user's configured environment; they are not included in this package. Audit-ready mode requires a separately deployed integrity authority; the skill package alone does not provide it.