Casefile

Overview Schema Source Related Servers Discussions Admin

add_project_entry

Record a project's decision, finding, artifact, or note in its case file. Keeps decisions and context persistent across AI agent sessions.

Instructions

Files an entry in a project's case: a decision, finding, artifact or note that concerns the project rather than one of its tasks.

The entry number counts inside the project, and TRK#7 addresses the entry from refs of any task or project case. Like a task entry filed by add_entry, a project entry stays as filed. A task token files project entries as it files task entries.

An empty title, or a reference to a missing entry, task or project, returns entry_fields_invalid naming the offending fields.

Input Schema

TableJSON Schema

Name	Required	Description
`key`	Yes	Project key, case-insensitive. An unknown key is refused with `project_not_found`
`body`	No	Entry body in markdown, stored and returned as is. It holds what a successor needs to continue; file contents and long outputs stay outside it, represented by a pointer and the gist
`refs`	No	References: task entries `TRK-42#12`, project entries `TRK#7`, tasks `TRK-7`, or a URL with a scheme (`https://…`, `file://…`). Anything else (`7`, `#7`, `docs/x.md`) is refused with `entry_fields_invalid`, as is an entry, task or project that does not exist; URLs are not checked
`type`	Yes	What the entry records about the project: - `decision` — an option chosen among several, with the reason; - `finding` — an established fact with its source; - `artifact` — a pointer to a result; - `note` — an entry that fits none of the types above. Summaries, questions, attempts, verdicts and remarks exist only in task cases
`title`	Yes	Entry title: its line in the case index of `get_project`. It states what happened, not how
`idempotency_key`	No	Retry key chosen by the caller, e.g. a UUID. A repeat with the same key and the same arguments returns the first result and creates nothing; the same key with other arguments is refused with `idempotency_key_reused`. A key is bound to the caller's token and kept for 24 hours

Output Schema

TableJSON Schema

Name	Required	Description
`no`	Yes	Entry number in the project's case; with the key it forms `TRK#7`
`seq`	Yes	Journal sequence number, usable as `after` of `wait_journal`
`author`	Yes
`created_at`	Yes
`project_key`	Yes

Schema Changelog

Changes observed during successful MCP inspections.

Changed1 schema field changedv0.8.1Sep 30, 2026

changedInput schema / properties / refs / description

Previous value: -"References: task entries `TRK-42#12`, project entries `TRK#7`, tasks `TRK-7`, URLs. An entry, task or project that does not exist is refused with `entry_fields_invalid`; URLs are not checked"New value: +"References: task entries `TRK-42#12`, project entries `TRK#7`, tasks `TRK-7`, or a URL with a scheme (`https://…`, `file://…`). Anything else (`7`, `#7`, `docs/x.md`) is refused with `entry_fields_invalid`, as is an entry, task or project that does not exist; URLs are not checked"

Addedv0.5.2Sep 25, 2026

TDQS

A4.1/5.0

Behavior4/5

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

Annotations already declare the write/safety profile (readOnlyHint=false, destructiveHint=false, idempotentHint=false). The description adds value beyond that: entries 'stay as filed' (immutability), the numbering is project-scoped, refs addressing via TRK#7, and the entry_fields_invalid failure mode. It does not restate the annotation fields, which is correct.

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?

Three short paragraphs, front-loaded with purpose then addressing then error behavior. Every sentence carries information, though the closing error paragraph partially duplicates what the schema's per-field descriptions already state.

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?

An output schema exists, so return values need not be explained, and the description covers purpose, addressing, immutability, and failure modes. What remains thin is explicit guidance on non-idempotent retries, though the idempotency_key param and idempotentHint=false together convey this.

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 description coverage is 100%, so all six parameters are already documented including the type enum and refs grammar. The description restates the entry types and error semantics rather than adding syntax not in the schema, so the baseline 3 holds.

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?

States a specific verb and resource ('files an entry in a project's case') and immediately scopes it against the sibling concept: an entry that 'concerns the project rather than one of its tasks'. An agent can distinguish this from add_entry without opening a schema.

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?

Gives the selecting condition (project-level vs task-level entry) and names the sibling add_entry as the task-side analogue. No explicit when-not-to-use or exclusion list, but the contrast with task entries is a clear routing signal.

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