Developer-tools MCP server

Citra

Version history of discovered tools, schemas, and declared agent authority.

View watchlist

DistributionDistribution: npm
Package@sylphx/citra
Latest analyzed version8.5.11
Current tool count4
Latest analysisNov 16, 2026
Package source

OBSERVED INTERFACE

Capability surface

Modelara’s observed declarations are shown separately from its derived effect classifications.

3 releases analyzed8.5.9 → 8.5.11Latest: 8.5.11
  • read2
  • write0
  • destructive0
  • unknown2

Unknown means Valyd could not determine the capability's effect confidently from the declared interface and static evidence. It is not treated as safe.

Package
@sylphx/citra@8.5.11
Registry
npm
Published
Nov 7, 2026
Integrity
sha512-wYgleZXqE9qip0wWwCVbER32UNSk/r4BwD+Nc6u4MYvQO4vApNT9LZfI15d/qB5hKVIasvf4/69W53jrXc/iSw==
Analyzer
modelara v0.1.8 · policy not recorded
Repository
https://github.com/SylphxAI/anymd

4 capabilities

CapabilityTypeDerived effectFirst observed in Valyd coverageLast changed
inspectsylphxai-citra:tool:inspect
Observed interface and hashes

Document and media evidence when Markdown is not enough. video_timeline (bounded scene/cue timeline) and render_frame (actual decoded frames), and cite_check (quote/location support, not semantic truth), are anymd Pro; PDF operations: inspect (page facts, metadata), render_page (PNG images), extract_regions (crop bounding boxes), ocr_pages / analyze_regions (configured OCR or vision provider), structure (JSON with document map, elements, geometry; profile quality|research adds trust and accessibility reports), compare (page-level diff of sources[0] vs sources[1]).

Input schema

{
  "$defs": {
    "Citation": {
      "additionalProperties": false,
      "properties": {
        "bounding_box": {
          "$ref": "#/$defs/CiteBoundingBox",
          "description": "Bottom-left PDF coordinates, finite and with positive area. No padding."
        },
        "id": {
          "type": [
            "string",
            "null"
          ]
        },
        "page": {
          "description": "One-based physical PDF page, not a printed page label. At most 20 distinct pages per request.",
          "minimum": 1,
          "type": "integer"
        },
        "quote": {
          "description": "Nonempty quote, at most 4096 UTF-16 units; request total at most 64000 units.",
          "maxLength": 4096,
          "minLength": 1,
          "type": "string"
        }
      },
      "required": [
        "quote",
        "page",
        "bounding_box"
      ],
      "type": "object"
    },
    "CiteBoundingBox": {
      "additionalProperties": false,
      "properties": {
        "bottom": {
          "type": "number"
        },
        "left": {
          "type": "number"
        },
        "right": {
          "type": "number"
        },
        "top": {
          "type": "number"
        }
      },
      "required": [
        "left",
        "bottom",
        "right",
        "top"
      ],
      "type": "object"
    },
    "CiteNormalization": {
      "enum": [
        "none",
        "whitespace_v1"
      ],
      "type": "string"
    },
    "InspectOperation": {
      "oneOf": [
        {
          "const": "video_timeline",
          "description": "Bounded local video timeline.",
          "type": "string"
        },
        {
          "const": "render_frame",
          "description": "Decode actual video frames at requested playback timestamps.",
          "type": "string"
        },
        {
          "const": "cite_check",
          "description": "Deterministic quote and location support, not semantic truth.",
          "type": "string"
        },
        {
          "const": "inspect",
          "description": "Page count, metadata, and per-page facts.",
          "type": "string"
        },
        {
          "const": "render_page",
          "description": "Render pages to PNG images.",
          "type": "string"
        },
        {
          "const": "extract_regions",
          "description": "Crop regions (bounding boxes) out of rendered pages.",
          "type": "string"
        },
        {
          "const": "ocr_pages",
          "description": "OCR pages with the configured provider.",
          "type": "string"
        },
        {
          "const": "analyze_regions",
          "description": "Send regions to a configured vision/region provider.",
          "type": "string"
        },
        {
          "const": "structure",
          "description": "Structured JSON read: document map, elements, geometry, trust and accessibility reports.",
          "type": "string"
        },
        {
          "const": "compare",
          "description": "Page-level text diff between two PDFs (sources[0] = before, sources[1] = after).",
          "type": "string"
        }
      ]
    },
    "OcrEngine": {
      "enum": [
        "auto",
        "vlm",
        "tesseract"
      ],
      "type": "string"
    },
    "PageNumber": {
      "minimum": 1,
      "type": "integer"
    },
    "PageSpecifier": {
      "anyOf": [
        {
          "items": {
            "$ref": "#/$defs/PageNumber"
          },
          "type": "array"
        },
        {
          "minLength": 1,
          "type": "string"
        }
      ]
    },
    "PdfEvidenceRegion": {
      "properties": {
        "bounding_box": {
          "$ref": "#/$defs/RegionBoundingBox"
        },
        "id": {
          "type": [
            "string",
            "null"
          ]
        },
        "padding": {
          "maximum": 200,
          "minimum": 0,
          "type": [
            "number",
            "null"
          ]
        },
        "page": {
          "minimum": 1,
          "type": "integer"
        }
      },
      "required": [
        "page",
        "bounding_box"
      ],
      "type": "object"
    },
    "PdfEvidenceSource": {
      "properties": {
        "pages": {
          "anyOf": [
            {
              "$ref": "#/$defs/PageSpecifier"
            },
            {
              "type": "null"
            }
          ]
        },
        "path": {
          "description": "Path to a local source (PDF for PDF operations; media for video operations). Provide exactly one of path or url (not both).",
          "minLength": 1,
          "type": [
            "string",
            "null"
          ]
        },
        "regions": {
          "items": {
            "$ref": "#/$defs/PdfEvidenceRegion"
          },
          "type": [
            "array",
            "null"
          ]
        },
        "url": {
          "description": "Explicit source URL. Provide exactly one of path or url (not both).",
          "minLength": 1,
          "type": [
            "string",
            "null"
          ]
        }
      },
      "type": "object"
    },
    "RegionBoundingBox": {
      "properties": {
        "bottom": {
          "minimum": 0,
          "type": "number"
        },
        "left": {
          "minimum": 0,
          "type": "number"
        },
        "right": {
          "minimum": 0,
          "type": "number"
        },
        "top": {
          "minimum": 0,
          "type": "number"
        }
      },
      "required": [
        "left",
        "bottom",
        "right",
        "top"
      ],
      "type": "object"
    },
    "TimelineSelection": {
      "additionalProperties": false,
      "properties": {
        "caption": {
          "default": false,
          "type": "boolean"
        },
        "end_ms": {
          "minimum": 0,
          "type": "integer"
        },
        "max_scenes": {
          "default": 20,
          "minimum": 0,
          "type": "integer"
        },
        "start_ms": {
          "default": 0,
          "minimum": 0,
          "type": "integer"
        }
      },
      "required": [
        "end_ms"
      ],
      "type": "object"
    }
  },
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "properties": {
    "citations": {
      "description": "cite_check only: 1–100 quotes with required page and bounding_box; exactly one PDF source.",
      "items": {
        "$ref": "#/$defs/Citation"
      },
      "maxItems": 100,
      "minItems": 1,
      "type": [
        "array",
        "null"
      ]
    },
    "expected_source_sha256": {
      "description": "cite_check and video operations: fail if the admitted source's SHA-256 differs from this hexadecimal value.",
      "maxLength": 64,
      "minLength": 64,
      "type": [
        "string",
        "null"
      ]
    },
    "include_image": {
      "type": [
        "boolean",
        "null"
      ]
    },
    "include_metadata": {
      "type": [
        "boolean",
        "null"
      ]
    },
    "languages": {
      "items": {
        "type": "string"
      },
      "type": [
        "array",
        "null"
      ]
    },
    "max_output_chars": {
      "maximum": 1000000,
      "minimum": 1000,
      "type": [
        "integer",
        "null"
      ]
    },
    "max_pages": {
      "maximum": 20,
      "minimum": 1,
      "type": [
        "integer",
        "null"
      ]
    },
    "max_pixels_per_page": {
      "maximum": 64000000,
      "minimum": 10000,
      "type": [
        "integer",
        "null"
      ]
    },
    "max_regions": {
      "maximum": 100,
      "minimum": 1,
      "type": [
        "integer",
        "null"
      ]
    },
    "normalization": {
      "anyOf": [
        {
          "$ref": "#/$defs/CiteNormalization"
        },
        {
          "type": "null"
        }
      ]
    },
    "ocr": {
      "anyOf": [
        {
          "$ref": "#/$defs/OcrEngine"
        },
        {
          "type": "null"
        }
      ],
      "description": "Local OCR engine for ocr_pages: auto, vlm, or tesseract. Omit to use the configured command provider."
    },
    "operation": {
      "$ref": "#/$defs/InspectOperation"
    },
    "profile": {
      "description": "structure only: fast (default), quality, or research (adds safety, trust, accessibility).",
      "type": [
        "string",
        "null"
      ]
    },
    "sample_pages": {
      "maximum": 20,
      "minimum": 1,
      "type": [
        "integer",
        "null"
      ]
    },
    "scale": {
      "maximum": 4,
      "minimum": 0.25,
      "type": [
        "number",
        "null"
      ]
    },
    "sources": {
      "items": {
        "$ref": "#/$defs/PdfEvidenceSource"
      },
      "type": "array"
    },
    "timeline": {
      "anyOf": [
        {
          "$ref": "#/$defs/TimelineSelection"
        },
        {
          "type": "null"
        }
      ],
      "description": "video_timeline only (required there): half-open playback window with required end_ms, at most ten minutes and twenty detected-cut scenes. caption only samples one frame per scene through a user-configured local-command adapter."
    },
    "timeout_ms": {
      "maximum": 300000,
      "minimum": 1000,
      "type": [
        "integer",
        "null"
      ]
    },
    "timestamps_ms": {
      "description": "render_frame only: playback timestamps in integer milliseconds (at most twenty). Returns decoded frame metadata with the actual decoded timestamp.",
      "items": {
        "minimum": 0,
        "type": "integer"
      },
      "type": [
        "array",
        "null"
      ]
    },
    "transcript": {
      "description": "video_timeline only: also attach timed ASR cues from preinstalled local Qwen3-ASR weights. Never downloads models. Default false.",
      "type": [
        "boolean",
        "null"
      ]
    }
  },
  "required": [
    "operation",
    "sources"
  ],
  "type": "object"
}
sha256:74fc2a73aa00c770549003316697c03a8af86078a9960da154e7299ba572c8ea

Output schema

Not exposed in the analyzed interface.

sha256:398b676e47d03a29016ee92fe378b8b4f1b3e965390c4c64ce78d27f79df74d1

Description hash: sha256:895d3cafc9fb79eb202f73a10ff0b0d1c37f37249b34f3e32da18a5f538ee3fb

Interface hash: sha256:fad7f06a78edd5c45b0cd60746c799bfe34c04390c86decd9bffc03d90d39241

Other observed Modelara metadata
{
  "id": "mcp:anymd/inspect",
  "source": {
    "type": "mcp",
    "server": "anymd",
    "operation": "inspect"
  },
  "declared": {
    "read_only": "unknown",
    "destructive": "unknown",
    "idempotent": "unknown",
    "open_world": "unknown"
  },
  "safety": {
    "idempotent": "unknown",
    "retry_safe": "false"
  },
  "metadata": {
    "mcp.annotations": "absent"
  },
  "fingerprint": "sha256:70d3b9ceea35f1eb796938d68d80328892bebe24098a278d2854b38051d050d8"
}
toolunknown8.5.9No change observed
outlinesylphxai-citra:tool:outline
Observed interface and hashes

Return a document heading tree with stable node ids, title paths, page/slide/chapter ranges and Markdown byte ranges. PDF bookmarks or detected headings, Word headings, PowerPoint slides, EPUB chapters and headings, HTML and Markdown. format json (default) or tree. Pass a node id to read to fetch that section.

Input schema

{
  "$defs": {
    "OcrEngine": {
      "enum": [
        "auto",
        "vlm",
        "tesseract"
      ],
      "type": "string"
    },
    "OcrSelection": {
      "anyOf": [
        {
          "type": "boolean"
        },
        {
          "$ref": "#/$defs/OcrEngine"
        }
      ],
      "description": "Retain the existing boolean MCP contract while adding named engines."
    },
    "TimelineSelection": {
      "additionalProperties": false,
      "properties": {
        "caption": {
          "default": false,
          "type": "boolean"
        },
        "end_ms": {
          "minimum": 0,
          "type": "integer"
        },
        "max_scenes": {
          "default": 20,
          "minimum": 0,
          "type": "integer"
        },
        "start_ms": {
          "default": 0,
          "minimum": 0,
          "type": "integer"
        }
      },
      "required": [
        "end_ms"
      ],
      "type": "object"
    }
  },
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "properties": {
    "format": {
      "description": "json (default) or tree.",
      "type": [
        "string",
        "null"
      ]
    },
    "images": {
      "description": "Embedded images: none (default) or refs. Match this option in node reads.",
      "type": [
        "string",
        "null"
      ]
    },
    "ocr": {
      "anyOf": [
        {
          "$ref": "#/$defs/OcrSelection"
        },
        {
          "type": "null"
        }
      ],
      "description": "OCR image-only pages; false by default. Accepts booleans or auto/vlm/tesseract."
    },
    "revisions": {
      "description": "Word revisions: markup (default), accept or reject. Match node reads.",
      "type": [
        "string",
        "null"
      ]
    },
    "source": {
      "description": "File path or http(s) URL.",
      "type": "string"
    },
    "timeline": {
      "anyOf": [
        {
          "$ref": "#/$defs/TimelineSelection"
        },
        {
          "type": "null"
        }
      ],
      "description": "Local video only: outline this bounded playback window from metadata (cuts, chapters, subtitles). No ASR, OCR or captions run."
    }
  },
  "required": [
    "source"
  ],
  "type": "object"
}
sha256:2348b8f05508e868e750bd4166e6440c2ca1ed0a92216c490b6ef106ec99225d

Output schema

Not exposed in the analyzed interface.

sha256:398b676e47d03a29016ee92fe378b8b4f1b3e965390c4c64ce78d27f79df74d1

Description hash: sha256:1a975d1cb4bba76d6f2f4d3f74bfd8e820bcd03c7368fc43780c3da68cab5ddc

Interface hash: sha256:4c74171800280f4828fa940d02fcc1609cefdb15b4329879746ced6b1dc659e4

Other observed Modelara metadata
{
  "id": "mcp:anymd/outline",
  "source": {
    "type": "mcp",
    "server": "anymd",
    "operation": "outline"
  },
  "declared": {
    "read_only": "unknown",
    "destructive": "unknown",
    "idempotent": "unknown",
    "open_world": "unknown"
  },
  "safety": {
    "idempotent": "unknown",
    "retry_safe": "false"
  },
  "metadata": {
    "mcp.annotations": "absent"
  },
  "fingerprint": "sha256:09d2eed9a047e8e572d058b777ba05b68d03cbf9bc5e5fd76104fe459bc4a7df"
}
toolunknown8.5.9No change observed
readsylphxai-citra:tool:read
Observed interface and hashes

Read any document as clean Markdown: PDF, Word (DOCX), PowerPoint (PPTX), Excel (XLSX/XLS/ODS), CSV, EPUB, HTML or a web URL, Markdown/text, images (metadata + OCR), audio/video (metadata, chapters, subtitles). Pages/slides/sheets carry <!-- page N --> style markers for citation. Long documents stop at max_tokens (default 20000) and end with a cursor to continue; choose pages with pages: "1-5,8". A directory returns its readable files.

Input schema

{
  "$defs": {
    "OcrEngine": {
      "enum": [
        "auto",
        "vlm",
        "tesseract"
      ],
      "type": "string"
    },
    "OcrSelection": {
      "anyOf": [
        {
          "type": "boolean"
        },
        {
          "$ref": "#/$defs/OcrEngine"
        }
      ],
      "description": "Retain the existing boolean MCP contract while adding named engines."
    },
    "TimelineSelection": {
      "additionalProperties": false,
      "properties": {
        "caption": {
          "default": false,
          "type": "boolean"
        },
        "end_ms": {
          "minimum": 0,
          "type": "integer"
        },
        "max_scenes": {
          "default": 20,
          "minimum": 0,
          "type": "integer"
        },
        "start_ms": {
          "default": 0,
          "minimum": 0,
          "type": "integer"
        }
      },
      "required": [
        "end_ms"
      ],
      "type": "object"
    }
  },
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "properties": {
    "cursor": {
      "description": "Continue a previous read with the cursor from its last line.",
      "type": [
        "string",
        "null"
      ]
    },
    "download_asr_model": {
      "default": null,
      "description": "Implies transcript. Downloads pinned Qwen3-ASR-1.7B Q8 weights (~2.2 GB) into the anymd cache. Downloads only when explicitly enabled. Legacy download_whisper_model is accepted as an alias. Default false.",
      "type": [
        "boolean",
        "null"
      ]
    },
    "images": {
      "default": null,
      "description": "Images embedded inside PDFs, DOCX, PPTX and EPUB files. \"refs\" (default) saves each meaningful raster image once to the anymd cache and marks its place in the Markdown as ![caption](absolute path) plus an <!-- image: WxH, page N --> comment, so you can open the file yourself; logos, icons and repeated headers are skipped. \"none\" leaves images out. Standalone image files are not affected.",
      "enum": [
        "refs",
        "none",
        null
      ],
      "type": [
        "string",
        "null"
      ]
    },
    "max_tokens": {
      "description": "Token budget (default 20000). Longer documents stop at a unit boundary and end with a cursor.",
      "minimum": 500,
      "type": [
        "integer",
        "null"
      ]
    },
    "node": {
      "description": "Read a section id returned by outline; repeat node with a continuation cursor.",
      "type": [
        "string",
        "null"
      ]
    },
    "ocr": {
      "anyOf": [
        {
          "$ref": "#/$defs/OcrSelection"
        },
        {
          "type": "null"
        }
      ],
      "description": "Local OCR: auto, vlm, or tesseract. Auto uses the installed doc-VLM engine and weights (`anymd setup ocr` installs both, an explicit opt-in), otherwise tesseract. vlm without that setup returns a one-line setup notice. No automatic downloads. Boolean false disables; true retains automatic OCR."
    },
    "pages": {
      "description": "Pages (PDF), slides, sheets, or chapters to read, e.g. \"1-5,8\". Default: all.",
      "type": [
        "string",
        "null"
      ]
    },
    "revisions": {
      "default": null,
      "description": "Tracked changes and comments in Word (DOCX) files. \"markup\" (default) shows them as CriticMarkup: {++inserted++}, {--deleted--}, {~~old~>new~~}, {==commented==}{>>Author (date): comment<<}, each change followed by its author and date. \"accept\" gives the text with every change accepted and \"reject\" the text with every change rejected, both without comments. A document with no tracked changes or comments reads the same either way.",
      "enum": [
        "markup",
        "accept",
        "reject",
        null
      ],
      "type": [
        "string",
        "null"
      ]
    },
    "source": {
      "description": "File path, http(s) URL, or directory. PDF, DOCX, PPTX, XLSX/XLS/ODS, CSV/TSV, EPUB, HTML, Markdown/text, images, audio/video, SRT/VTT. A directory returns the list of readable files.",
      "minLength": 1,
      "type": "string"
    },
    "timeline": {
      "anyOf": [
        {
          "$ref": "#/$defs/TimelineSelection"
        },
        {
          "type": "null"
        }
      ],
      "description": "Local video only: read this bounded playback window as deterministic timeline sections. end_ms is required; the window is half-open, at most ten minutes and twenty scenes."
    },
    "transcript": {
      "default": null,
      "description": "Transcribe audio/video locally with bundled transcribe-cpp and Qwen3-ASR-1.7B Q8. Needs preinstalled SHA-256 pinned weights (~2.2 GB), unless download_asr_model is true. Audio stays local. Default false.",
      "type": [
        "boolean",
        "null"
      ]
    }
  },
  "required": [
    "source"
  ],
  "type": "object"
}
sha256:80b92271c6321535dcbf1826af442f5d29ae01261f707e1a08fe6606bddfc3df

Output schema

Not exposed in the analyzed interface.

sha256:398b676e47d03a29016ee92fe378b8b4f1b3e965390c4c64ce78d27f79df74d1

Description hash: sha256:3bf511227238e3609e8a4af0efdbf69628750b853d6dd63bd35fb64d986d886e

Interface hash: sha256:5ed703f06b9b19dc43b25c230f8ef53d4f9b7699fcc738a9bdcdef1b1b438372

Other observed Modelara metadata
{
  "id": "mcp:anymd/read",
  "source": {
    "type": "mcp",
    "server": "anymd",
    "operation": "read"
  },
  "declared": {
    "read_only": "unknown",
    "destructive": "unknown",
    "idempotent": "unknown",
    "open_world": "unknown"
  },
  "safety": {
    "idempotent": "unknown",
    "retry_safe": "false"
  },
  "metadata": {
    "mcp.annotations": "absent"
  },
  "fingerprint": "sha256:659ae1ca98679c8d7ce67c0d5c8aedeac47f62b796ff16b1d4e4547a3a47737d"
}
toolread8.5.9No change observed
searchsylphxai-citra:tool:search
Observed interface and hashes

Search documents for text: one file, many files, whole directories (recursive, .gitignore aware), or URLs, across every format read supports. Returns each hit as file + page/slide/sheet + a snippet with the match in bold. mode auto (default) finds the exact phrase and falls back to BM25-ranked passages when there is none; literal or ranked force one. Narrow directories with glob, e.g. "*.pdf".

Input schema

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "properties": {
    "case_sensitive": {
      "default": null,
      "type": [
        "boolean",
        "null"
      ]
    },
    "context_chars": {
      "description": "Snippet context characters on each side (default 80).",
      "maximum": 1000,
      "minimum": 0,
      "type": [
        "integer",
        "null"
      ]
    },
    "glob": {
      "description": "Only search files matching this glob inside directories, e.g. \"*.pdf\" or \"reports/**\".",
      "type": [
        "string",
        "null"
      ]
    },
    "max_results": {
      "description": "Maximum hits to return (default 20).",
      "maximum": 500,
      "minimum": 1,
      "type": [
        "integer",
        "null"
      ]
    },
    "mode": {
      "description": "auto (default): exact phrase, falling back to ranked passages when nothing matches. literal: exact phrase only. ranked: BM25 over the query words.",
      "type": [
        "string",
        "null"
      ]
    },
    "query": {
      "description": "Text to find.",
      "minLength": 1,
      "type": "string"
    },
    "sources": {
      "default": [],
      "description": "Files, directories (searched recursively, .gitignore respected), or URLs. Default: the current directory.",
      "items": {
        "type": "string"
      },
      "type": "array"
    },
    "whole_word": {
      "default": null,
      "type": [
        "boolean",
        "null"
      ]
    }
  },
  "required": [
    "query"
  ],
  "type": "object"
}
sha256:0ca553f3479d128dbd0c7c1723bf7ac9fa5d00e6420bd2f2f9ab5b256346a69d

Output schema

Not exposed in the analyzed interface.

sha256:398b676e47d03a29016ee92fe378b8b4f1b3e965390c4c64ce78d27f79df74d1

Description hash: sha256:60da1cc39d86e86c7c1e60603ee35bc0079d4a69c08ffc449c2395e0e1996e65

Interface hash: sha256:7c3a23ddaa3501a1a772495b54fce1109baf1aab7fb5f8d7f94f2a334c5c5bf9

Other observed Modelara metadata
{
  "id": "mcp:anymd/search",
  "source": {
    "type": "mcp",
    "server": "anymd",
    "operation": "search"
  },
  "declared": {
    "read_only": "unknown",
    "destructive": "unknown",
    "idempotent": "unknown",
    "open_world": "unknown"
  },
  "safety": {
    "idempotent": "unknown",
    "retry_safe": "false"
  },
  "metadata": {
    "mcp.annotations": "absent"
  },
  "fingerprint": "sha256:b90cddc9efaf3b0ba2f3c2c4c39cb3ad32898c3d71d8fba4ae430dd7b6908f35"
}
toolread8.5.9No change observed

STATIC COMPARISONS

Release history

2 release transitions
  1. 0 added0 removed0 schema0 write0 destructive0 unknown

    Nov 16, 2026· Automated static analysis

  2. 0 added0 removed0 schema0 write0 destructive0 unknown

    Nov 16, 2026· Automated static analysis