Skip to main content
Glama
phasespace-labs

palinode

by phasespace-labs
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Hybrid

palinode_save

Save facts, decisions, insights, or project updates as persistent, searchable markdown memories with optional citations, claims, and metadata. Prevents data loss across sessions and enables later retrieval via search.

Instructions

Save a memory (fact, decision, insight, project update) worth keeping across sessions. Requires exactly one of type or ps=true. On timeout the save may still have committed — palinode_search a distinctive phrase before retrying, or you'll duplicate it.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
psNoShorthand for type=ProjectSnapshot (the CLI `--ps` flag). If true, omit `type`; any other type value errors.
coreNoIf true, this memory is always injected at session start (core memory).
slugNoOptional URL-safe filename slug (auto-generated if omitted)
typeNoMemory type. Required unless `ps=true` is given.
titleNoHuman-readable title, used in list/search displays.
claimsNoBinds each claim to the source span justifying it. Read back via palinode_blame(claims=true).
sourceNoSource surface that created this memory.
contentYesThe memory content to save (markdown supported)
projectNoProject slug shorthand — 'palinode' becomes entity 'project/palinode'.
sourcesNoCitation anchors for passages this memory quotes.
entitiesNoRelated entity refs e.g. ['person/alice', 'project/alpha']
metadataNoAdditional frontmatter fields to merge into the saved memory.
priorityNoHuman-assigned memory priority (1–5). Stored as `priority` frontmatter; missing means normal (3).
backed_byNoRefs (category/slug) that support/back this memory (evidence links).
epistemicNoKind of claim: fact=observed, inference=derived, open_question=unresolved, unverified=asserted but unchecked. Omit to leave unmarked — unmarked is NOT fact.
confidenceNoConfidence in this memory's accuracy (0.0-1.0).
contradictsNoRefs (category/slug) this memory conflicts with; neither wins — surfaced for review.
external_refsNoSDLC object references such as github_pr or jira_issue.
update_policyNoRe-save to same slug: 'append' keeps the existing body and adds below it; 'replace' overwrites and marks a living doc. Sticky.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed1 schema field changedv0.20.1
    • changedInput schema / properties / update_policy / description
      Previous value: -"Save behavior: append episodic memory or replace a living document."New value: +"Re-save to same slug: 'append' keeps the existing body and adds below it; 'replace' overwrites and marks a living doc. Sticky."
  2. Changed13 schema fields changedv0.10.0
    • changedInput schema / properties / claims / description
      Previous value: -"Claim-level source anchors: list of {text, source_id, span:{quote, quote_hash}} bindings resolving each claim to the source span that justifies it. claim_id is derived on save; read back via palinode_blame with claims=true."New value: +"Binds each claim to the source span justifying it. Read back via palinode_blame(claims=true)."
    • changedInput schema / properties / claims / items / properties / anchor_id / description
      Previous value: -"Optional opaque pointer within a large source (interop; nullable)."New value: +"Optional pointer within a large source."
    • changedInput schema / properties / claims / items / properties / claim_id / description
      Previous value: -"Optional stable claim id; content-addressed, derived on save if omitted."New value: +"Optional; derived on save."
    • changedInput schema / properties / claims / items / properties / source_id / description
      Previous value: -"Path under the memory dir of the source that justifies the claim (a sources[].ref)."New value: +"A sources[].ref that justifies the claim."
    • changedInput schema / properties / claims / items / properties / span / properties / quote / description
      Previous value: -"The exact passage in the source that justifies the claim."New value: +"The justifying passage in the source."
    • changedInput schema / properties / claims / items / properties / span / properties / quote_hash / description
      Previous value: -"Optional integrity hash; computed on save if omitted."New value: +"Optional; computed on save."
    • changedInput schema / properties / epistemic / description
      Previous value: -"Epistemic marker: 'fact' (observed/verified), 'inference' (derived, lower trust), 'open_question' (unresolved), or 'unverified' (asserted but not checked). Omit to leave the memory unmarked (no claim is made — not treated as fact)."New value: +"Kind of claim: fact=observed, inference=derived, open_question=unresolved, unverified=asserted but unchecked. Omit to leave unmarked — unmarked is NOT fact."
    • changedInput schema / properties / project / description
      Previous value: -"Project slug shorthand — e.g. 'palinode' becomes entity 'project/palinode'.  Pairs with `palinode_session_end`'s `project` field for consistent project tagging across save and session-end."New value: +"Project slug shorthand — 'palinode' becomes entity 'project/palinode'."
    • changedInput schema / properties / ps / description
      Previous value: -"Shorthand for type=ProjectSnapshot — matches the CLI `--ps` flag and the `/ps` slash command. If true, `type` may be omitted (or set to ProjectSnapshot redundantly); other type values conflict and error."New value: +"Shorthand for type=ProjectSnapshot (the CLI `--ps` flag). If true, omit `type`; any other type value errors."
    • changedInput schema / properties / sources / description
      Previous value: -"Source-citation anchors: list of {ref, quote, quote_hash} for passages this memory cites."New value: +"Citation anchors for passages this memory quotes."
    • changedInput schema / properties / sources / items / properties / quote / description
      Previous value: -"The exact passage cited from the source."New value: +"The exact passage cited."
    • changedInput schema / properties / sources / items / properties / quote_hash / description
      Previous value: -"Optional integrity hash; computed on save if omitted."New value: +"Optional; computed on save."
    • changedInput schema / properties / title / description
      Previous value: -"Optional human-readable title.  Stored in frontmatter and used in list/search displays."New value: +"Human-readable title, used in list/search displays."
  3. First observedv0.9.5

TDQS

A4.1/5.0
Behavior4/5

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

The description adds an important behavioral caveat beyond the annotations: a timeout may still commit the save, so retrying blindly duplicates memories. It also states the type/ps one-of requirement. This meaningfully supplements the idempotentHint=false annotation.

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

Conciseness5/5

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

Two sentences with no filler. The purpose is front-loaded, and the second sentence carries the critical failure-handling guidance. Every sentence earns its place.

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

Completeness3/5

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

For a complex 19-parameter write tool with no output schema, the description is rather thin. It covers the one-of constraint and timeout behavior, but does not explain what the tool returns, the update_policy re-save semantics, or what a successful save looks like. The schema covers parameters, but an agent still lacks a full picture of the tool's side effects and return value.

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 the schema already documents all parameters. The description adds the cross-parameter "exactly one of type or ps=true" rule, but the ps schema description already conveys most of this. It provides no additional value for the remaining parameters.

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 states a specific verb and resource: "Save a memory" with concrete categories (fact, decision, insight, project update). This clearly distinguishes the tool from its read/search/list siblings even without opening schemas.

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?

The description gives clear context for when to use the tool (anything worth keeping across sessions) and includes a concrete exclusion/retry rule: search before retrying after timeout to avoid duplicates. It does not explicitly enumerate alternative tools, but the usage context is clear enough.

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

Deploy Server

Other Tools