docs(ai),liveness: open-edition honesty for the agent capability - #2812
Merged
Conversation
Three fixes from the whole-platform agent-capability audit — all "declared ≠ delivered on the open edition" cleanups, no behavior change: 1. aggregate_data over-claim — Natural Language Queries said your own AI gets `query_records` / `get_record` / `aggregate_data` via the open `@objectstack/mcp`. The open native HTTP surface registers 9 tools and NONE is `aggregate_data`; `query_records` has no aggregation args. `aggregate_data` is a cloud data tool. Doc now lists the real open read tools and fences aggregation to the cloud runtime. 2. "Skill" terminology collision — `docs/ai/skills.mdx` (authoring SKILL.md modules for coding assistants, open, skills.sh) and `defineSkill` / SkillSchema in `agents.mdx` (runtime capability bundles on ask/build, cloud) are unrelated concepts sharing one word, both under /docs/ai. Added cross-linked disambiguation callouts to both pages. 3. liveness cites cloud code as framework evidence — agent/skill/tool/action liveness files point `evidence` at `packages/services/service-ai/...`, which is a stale untracked build artifact; the real consumer is the closed cloud `@objectstack/service-ai`. Each `_note` now states this, so the open/cloud line is clear to anyone auditing what the open edition actually enforces. Liveness gate green (staleEvidence stays a warning, not a failure); Callout variants/anchors/tag-balance verified. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
github-actions
Bot
added
documentation
Contributor
📓 Docs Drift CheckThis PR changes 1 package(s): 94 hand-written doc(s) reference the affected code and may need an implementation-accuracy re-verification:
|
This was referenced Jul 11, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Three "declared ≠ delivered on the open edition" cleanups surfaced by a whole-platform agent-capability audit. No behavior change — docs + liveness
_noteannotations only.1.
aggregate_dataover-claim (Natural Language Queries)The doc told readers their own AI gets
query_records/get_record/aggregate_datavia the open@objectstack/mcp. But the open native HTTP surface (mcp-http-tools.ts) registers exactly 9 tools —list_objects,describe_object,query_records,get_record,create_record,update_record,delete_record,list_actions,run_action— and none isaggregate_data;query_recordshas no aggregation arguments.aggregate_datais a cloud data tool (cloud/.../tools/data-tools.ts); in framework the name appears only in a read-only classification set (mcp-server-runtime.ts:54), never as a registered tool. Fixed: the doc lists the real open read tools and fences aggregation to the cloud runtime with a callout.2. "Skill" terminology collision
Two unrelated concepts share the word "skill," both under
/docs/ai:skills.mdx/skills-reference.mdx) =SKILL.mdknowledge modules (skills.sh) that teach a coding assistant to write metadata — open, dev-time, never run.defineSkill/SkillSchemainagents.mdx) = runtime capability bundles attached to theask/buildplatform agents — cloud runtime.agents.mdxeven used both senses within a few lines. Added cross-linked disambiguation callouts to both pages.3. Liveness cites cloud code as framework evidence
The
agent/skill/tool/actionliveness ledgers pointevidenceatpackages/services/service-ai/src/...— but that framework tree is a stale, untracked build artifact (git ls-filesreturns nothing; nosrc/, nopackage.json). The real consumer is the closed cloud@objectstack/service-ai. So the open framework's own ledger reads as if it locally enforces an agent runtime it doesn't contain — the single most confusing artifact for anyone auditing "what does open actually do." Each file's_notenow states the evidence lives in cloud/EE and these props arelivebecause that cloud runtime consumes them.Verification
check:livenessexit 0, "all governed-type properties classified";staleEvidenceis a warning that never counted toward failure, so no regression).type="warn"is an already-used variant; anchor target confirmed; Callout tags balanced in all edited files; all four liveness JSONs re-validated.The audit's larger finding — that the platform's agent capability is bifurcated (open = MCP + authoring schemas + knowledge; cloud/EE = the entire execution runtime, ask/build, conversation store, chat backend) — is left as-is: the open/cloud line in
content/docs/ai/index.mdxalready draws it correctly. This PR only removes the three concrete honesty gaps.🤖 Generated with Claude Code