Skip to main content
Glama

Check citations of the African Commission and Court in a text

banjul_check

Checks citations in a draft against African Commission and Court decisions, verifying case numbers, names, paragraphs, and quotations to flag missing, ambiguous, or mismatched references.

Instructions

Corpus: decisions of the African Commission on Human and Peoples' Rights (from its library) and English judgments, rulings, orders and separate opinions of the African Court on Human and Peoples' Rights (from AJUDATA), split into their numbered paragraphs. Scanned decisions were read by OCR: check the PDF before quoting. Finds the citations in a text (communication numbers "Communication 245/02", application numbers "Application No. 005/2013", names "X v State", "X, cited above") and checks each: does the number exist and match the name, which decision of the case is meant (by date, year or kind), does the pinpointed paragraph exist, and does a quotation next to the citation match the decision word for word. Status per citation: ok, warn (check it), bad (not found), amb (ambiguous), out (outside this database, e.g. ECOWAS Court). Use it on your own draft answer before giving it.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "text": "<text>"
    },
    "name": "banjul_check"
  }
}

Input Schema

NameRequiredDescriptionDefault
textYesThe text to check (up to about 30,000 characters).
compare_quotesNo

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv1.0.0

TDQS

A4.1/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the full burden and does substantial work: it discloses the corpus boundaries, an OCR reliability caveat, and the full status vocabulary (ok, warn, bad, amb, out) with meanings, plus the note that 'out' covers courts outside the database such as ECOWAS. What it omits is whether anything about the submitted text is retained, and there is no statement of latency or size limits beyond the schema's character cap.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The prose is dense but purposeful, with no filler sentences; the corpus definition, the checks performed, the status codes and the usage instruction all earn their place. The one structural weakness is ordering: the corpus inventory occupies the opening position, so the actual action verb ('Finds the citations in a text ... and checks each') arrives after two sentences of setup.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a two-parameter tool with no output schema, the description supplies the return semantics via the status vocabulary, which is the main thing an agent needs to interpret results. What remains thin is the granularity of a non-'ok' result: it does not say whether a warning or bad citation comes back with the matched decision identifier and paragraph, so the agent knows what it will have to repair.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 50%: `text` is documented in the schema itself, but `compare_quotes` has no description. The description implies quote comparison ('does a quotation next to the citation match the decision word for word') but never names the switch or explains what disabling it does — whether quotes are then skipped or reported unchecked. That is partial compensation, so the baseline-3 level applies.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description names a specific verb+object pair: it extracts citations from a text and verifies each one against a defined corpus (African Commission library + AJUDATA African Court judgments split into numbered paragraphs). It also enumerates the citation forms it recognises (communication numbers, application numbers, case names, 'cited above'), which makes the tool unmistakably a verification step rather than a search or retrieval tool like its siblings.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

It gives an explicit use case — 'Use it on your own draft answer before giving it' — which tells the agent exactly when this belongs in a workflow. It also flags a precondition ('check the PDF before quoting' for OCR'd scanned decisions). It does not explicitly route to siblings (banjul_search, banjul_get_decision) for follow-up retrieval, which keeps it short of a 5.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.