@ssheleg/agent-stack 0.23.1 → 0.23.2

package/CHANGELOG.md

@@ -1,3 +1,36 @@
+## v0.23.2 — a path that resolves only from the neighbour's directory, and the description with no room left
+
+Family audit 2026-09-06 (wave AUDIT-WAVE-0906) brought three findings for this member.
+Two changed files; the third changed a board instead.
+
+**Two references pointed at `references/kv-cache.md` from directories where it does not
+exist.** `agent-evals/references/otel-genai.md:137` and
+`agent-interop/references/mcp-scale.md:92` both named the bare path; the file lives only
+in sibling `agent-orchestrator`. The prose around each named the sibling, so a reader
+could recover the target — but the path as written resolves to nothing from either
+skill's own root, and a bare `references/x.md` means *this skill's file* by the
+validator's own rule (`test/validate.py:263-267`). The check never saw them because it
+reads SKILL.md alone; these sat in reference files nothing scans. Both now read
+`../agent-orchestrator/references/kv-cache.md`, which resolves from each skill's root —
+and the leading `../` keeps the validator's lookbehind from reading it as a claim on a
+local file.
+
+**`agent-interop`'s description sat at exactly 970/970 — the working limit with zero
+headroom** — so the next trigger, the next boundary clause, any edit at all would have
+fought the gate before doing its work. Eleven characters of connective prose trimmed
+("another agent" → "an agent"; two list-internal "and"s → commas): now **959/970**, with
+every quoted trigger phrase untouched — **14 before, 14 after**, English and Russian,
+counted by re-folding the YAML scalar rather than by eyeballing the diff.
+
+**The audit also asked for a debulk of `agent-orchestrator` and `agent-evals`, and that
+was declined rather than done blind.** Re-measured at this commit: **4655/4750** (98.0%)
+and **4631/4750** (97.5%), both `0 GAP`. The family precedent (B-141 class) is that
+manufacturing headroom with no pending edit is negative-value work: room bought today is
+spent by nobody, and a trim chosen without the edit it serves cuts on the wrong line.
+Filed as board row **B-135** — the debulk belongs to the release that next edits these
+bodies, beside B-129's standing record that the durable remedy for `agent-orchestrator`
+is a split, not a trim.
+
 ## v0.23.1 — the registry card stopped being ten releases stale
 
 `SKILL-CARD.md` carries the fields Anthropic's Skills-for-enterprise guidance asks every

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ssheleg/agent-stack",
-  "version": "0.23.1",
+  "version": "0.23.2",
   "scripts": {
     "test": "python3 test/validate.py && python3 test/plant_guard_test.py && node test/installer_test.js"
   },

package/plugins/agent-stack/.claude-plugin/plugin.json

@@ -3,7 +3,7 @@
   "name": "agent-stack",
   "displayName": "Agent Stack",
   "description": "Four skills: agent-orchestrator — tool-calling loops, pipelines with checkpoints, provider routing with fallback, memory architecture, plus the wallet side of reselling LLM access; agent-evals — run/trace/thread evals, LLM judges, and fixtures grown from production; agent-interop — MCP servers and clients, A2A agent cards, the MCP Registry, and gateways; agent-harness — system prompts, tool shaping, workflow-vs-agent, and auditing an agent system.",
-  "version": "0.23.1",
+  "version": "0.23.2",
   "author": {
     "name": "ssheleg",
     "url": "https://x.com/sshlg93"

package/plugins/agent-stack/skills/agent-evals/references/otel-genai.md

@@ -134,7 +134,7 @@ per-modality `text.*` / `image.*` / `audio.*` splits including
 
 **A cost computed from `input_tokens + output_tokens` alone is wrong in both directions.** It
 bills cache reads at full price — they are the cheap ones — and it misses reasoning tokens and
-cache writes entirely. `references/kv-cache.md` in `agent-orchestrator` is the other half of
+cache writes entirely. `../agent-orchestrator/references/kv-cache.md` is the other half of
 this: the cache read is the case worth getting right, because at scale it is most of the
 traffic.
 

package/plugins/agent-stack/skills/agent-interop/SKILL.md

@@ -2,11 +2,11 @@
 name: agent-interop
 description: >-
   Use when an agent must talk to something outside its own process — building or consuming an
-  MCP server, exposing or calling another agent over A2A, publishing to the MCP Registry, or
+  MCP server, exposing or calling an agent over A2A, publishing to the MCP Registry, or
   putting a gateway in front of agent traffic. Carries the MCP 2026-07-28 wire surface and what it
   deprecated (server/discover, stateless per-request _meta, elicitation in form and URL mode,
   subscriptions/listen; sampling, roots, logging and dynamic client registration going),
-  A2A 1.0 agent cards, task states and three bindings, registry namespaces and server.json, tool
+  A2A 1.0 agent cards, task states, three bindings, registry namespaces, server.json, tool
   federation, and what a gateway must do that an API gateway does not. Triggers - "MCP server",
   "MCP client", "A2A", "agent card", "agent interoperability", "MCP registry", "server.json",
   "agentgateway", "tool federation", "MCP-сервер", "карточка агента", "интероперабельность

package/plugins/agent-stack/skills/agent-interop/references/mcp-scale.md

@@ -88,8 +88,8 @@ tokens than the definitions you so carefully removed.
 mitigations below were approximating: `defer_loading` keeps deferred definitions out of the
 system-prompt prefix and appends discovered ones inline as `tool_reference` blocks, leaving the
 cached prefix untouched (at least one tool must stay non-deferred — all-deferred is a 400). The
-economics, the arithmetic and the other invalidators are in `agent-orchestrator`'s
-`references/kv-cache.md`.
+economics, the arithmetic and the other invalidators are in
+`../agent-orchestrator/references/kv-cache.md`.
 
 Three mitigations, in the order they are usually right: