mastra-ai/mastra · error · Error

Thread knowledge scope requires resource and org ancestors

Error message

Thread knowledge scope requires resource and org ancestors

What it means

Scopes are strictly hierarchical: a thread-scoped entry ('thread:<id>') is only meaningful beneath its resource and org ancestors. canonicalizeKnowledgeScope therefore requires that any scope containing a thread entry also contains the corresponding resource and org entries.

Source

Thrown at packages/core/src/storage/domains/knowledge/base.ts:295

  const entriesByLevel = new Map<KnowledgeScopeLevel, string>();
  for (const entry of scope) {
    const separator = entry.indexOf(':');
    const level = entry.slice(0, separator) as KnowledgeScopeLevel;
    const id = entry.slice(separator + 1);
    if (separator <= 0 || !id || id.includes('\u001f') || SCOPE_ORDER[level] === undefined) {
      throw new Error(`Invalid knowledge scope entry: ${entry}`);
    }
    const existing = entriesByLevel.get(level);
    if (existing && existing !== entry) {
      throw new Error(`Knowledge scope contains multiple ${level} entries`);
    }
    entriesByLevel.set(level, entry);
  }
  if (entriesByLevel.size === 0) {
    throw new Error('Knowledge scope cannot be empty');
  }
  if (entriesByLevel.has('thread') && (!entriesByLevel.has('resource') || !entriesByLevel.has('org'))) {
    throw new Error('Thread knowledge scope requires resource and org ancestors');
  }
  if (entriesByLevel.has('resource') && !entriesByLevel.has('org')) {
    throw new Error('Resource knowledge scope requires an org ancestor');
  }

  const unique = [...new Set(scope)];
  unique.sort((a, b) => {
    const aLevel = a.slice(0, a.indexOf(':')) as KnowledgeScopeLevel;
    const bLevel = b.slice(0, b.indexOf(':')) as KnowledgeScopeLevel;
    const aOrder = SCOPE_ORDER[aLevel] ?? Number.MAX_SAFE_INTEGER;
    const bOrder = SCOPE_ORDER[bLevel] ?? Number.MAX_SAFE_INTEGER;
    return aOrder - bOrder || a.localeCompare(b);
  });
  return unique;
}

export function knowledgeScopeKey(scope: KnowledgeScope): string {
  return canonicalizeKnowledgeScope(scope).join('\u001f');

View on GitHub (pinned to 75dd419e61)

Solutions

  1. Always pass the full ancestor chain: ['org:<o>', 'resource:<r>', 'thread:<t>'].
  2. Use expandKnowledgeScope(context, 'thread') to derive a properly ancestor-complete thread scope instead of hand-building it.
  3. When narrowing, keep all wider-level entries and only swap the narrowest one.
  4. Validate presence of every ancestor level in a helper before calling knowledge APIs.

Example fix

// before
const scope = [`thread:${threadId}`];
// after
const scope = [`org:${orgId}`, `resource:${resourceId}`, `thread:${threadId}`];
Defensive patterns

Strategy: validation

Validate before calling

function hasThreadAncestors(scope: string[]): boolean {
  const levels = new Set(scope.map(e => e.slice(0, e.indexOf(':'))));
  if (levels.has('thread')) return levels.has('resource') && levels.has('org');
  if (levels.has('resource')) return levels.has('org');
  return levels.size > 0;
}

Prevention

When it happens

Trigger: Calling any knowledge API with scope ['org:a', 'thread:t1'] (missing resource) or ['thread:t1'] alone; dropping the resource entry when narrowing a scope from resource to thread level.

Common situations: Constructing thread scope from just a threadId; copying a scope array and deleting the resource line for 'narrowing'; passing expandKnowledgeScope output of the wrong level back in as raw context.

Related errors


AI-assisted analysis of mastra-ai/mastra@75dd419e61 (2026-08-30). Data as JSON: /api/errors/cb66d08349c9a300. Report an issue: GitHub.