diff --git a/bin/gstack-render-claude.sh b/bin/gstack-render-claude.sh
index e7f45ed93..c7b06c4cd 100644
--- a/bin/gstack-render-claude.sh
+++ b/bin/gstack-render-claude.sh
@@ -64,9 +64,9 @@ gstack_claude_overlay() {
 # for a saved non-full claude_loadout, else empty. A malformed value is
 # ignored with a note, so the full skills are served.
 gstack_claude_loadout_arg() {
-  local value
+  local CONFIG_BIN="$1" value
   _LOADOUT_ARG=""
-  value="$("$1" get claude_loadout 2>/dev/null || true)"
+  value="$("$CONFIG_BIN" get claude_loadout 2>/dev/null || true)"
   case "$value" in
     ''|full) return 0 ;;
   esac
diff --git a/capy/skill/gstack/SKILL.md b/capy/skill/gstack/SKILL.md
index cff559940..25a087975 100644
--- a/capy/skill/gstack/SKILL.md
+++ b/capy/skill/gstack/SKILL.md
@@ -30,7 +30,7 @@ With `GSTACK_CAPY=1`, a workflow started in an unattended or spawned session che
 
 For Capy running GPT-6 Astra, read `${CODEX_HOME:-$HOME/.codex}/skills/gstack/SKILL.md` for routing, then `${CODEX_HOME:-$HOME/.codex}/skills/gstack-<skill>/SKILL.md` for the selected workflow. These are the generated GPT-6 Astra instructions, not the Claude source files.
 
-For Claude Code, read `~/.claude/skills/gstack/SKILL.md` for routing and `~/.claude/skills/<skill>/SKILL.md` for the selected workflow. When launching a Claude Code session, tell it to load gstack and name the intended workflow. For Capy running a Claude model, the Claude instruction files can be read directly with the Capy adaptations below; that does not turn Capy into the Claude Code harness. Add `--model <id>` (your model id from the thread's model line) to every `scripts/install.sh` run: when it prints `SKILL_ROOT: <dir>`, read `<dir>/gstack/SKILL.md` and `<dir>/<skill>/SKILL.md` instead and source the printed `GSTACK_LAUNCH_FILE` (`docs/loadouts.md`; only Opus 5.5 gets a smaller tree).
+For Claude Code, read `~/.claude/skills/gstack/SKILL.md` for routing and `~/.claude/skills/<skill>/SKILL.md` for the selected workflow. When launching a Claude Code session, tell it to load gstack and name the intended workflow. For Capy running a Claude model, the Claude instruction files can be read directly with the Capy adaptations below; that does not turn Capy into the Claude Code harness. Pass your model id as `--model <id>` to `scripts/install.sh`; when it prints `SKILL_ROOT: <dir>`, read skills from `<dir>` instead (`docs/loadouts.md`).
 
 If no specific skill was requested, use the router's task-to-skill mapping. Read only the selected workflow and its referenced files. For other models, use the closest applicable instruction profile without claiming a model-specific profile was generated or evaluated for them.
 
diff --git a/test/fixtures/context-budget.json b/test/fixtures/context-budget.json
index 9f26e133f..8ea2838e3 100644
--- a/test/fixtures/context-budget.json
+++ b/test/fixtures/context-budget.json
@@ -9,7 +9,7 @@
     "browser-skills/hackernews-frontpage": 371,
     "canary": 7738,
     "captain": 6551,
-    "capy/skill/gstack": 5136,
+    "capy/skill/gstack": 5690,
     "careful": 1239,
     "codex": 9890,
     "context-restore": 6287,