audit-tools 0.32.5 → 0.32.7

package/skills/audit-code/opencode-command-template.txt

@@ -3,11 +3,10 @@
 Use `audit-code next-step` as the primary interface to the audit workflow.
 
 1. Run `audit-code next-step` directly when shell access is available.
-2. If MCP is your only available interface, call `start_audit` or `continue_audit`; both return the same one-step contract.
-3. Read `prompt_content` in the response and follow it.
-4. When a step completes (not blocked), run `audit-code next-step` again or call `continue_audit` as the compatibility adapter.
-5. Stop when the step instructions say to stop.
+2. Read `prompt_content` in the response and follow it.
+3. When a step completes (not blocked), run `audit-code next-step` again.
+4. Stop when the step instructions say to stop.
 
 Use the `task` tool or equivalent for subagent dispatch when the step tells you to fan out review work.
 
-If neither shell access nor `start_audit` is available, stop and report that no next-step interface is connected.
+If shell access to `audit-code next-step` is unavailable, stop and report that no next-step interface is connected.

package/skills/remediate-code/remediate-code.prompt.md

@@ -55,9 +55,13 @@ Every `next-step` call is also the **capability handshake**: report what you can
 dispatch to *right now* so the backend sizes remediation waves and per-worker
 context to your real model instead of a conservative 32k floor. Report:
 
-- `--host-can-dispatch-subagents` / `--host-max-concurrent` — whether you can run
-  callable subagents and how many in parallel. Without them the backend assumes
-  serial dispatch.
+- `--host-can-dispatch-subagents` — whether you can run callable subagents at all
+  (via the `Agent`/`task` tool). Without it the backend runs remediation serially.
+  Do **not** report a parallel-worker count: the backend owns concurrency (its
+  token-budget gate plus any hard host cap it can detect, e.g. Codex
+  `[agents].max_threads`) and runs uncapped when it detects none. An operator with
+  a genuine fixed machine limit may pass `--host-max-concurrent N` as an optional
+  override; it is never a value to guess.
 - `--host-models` — an ordered JSON array (lowest rank first) of the models you
   can dispatch workers to *right now*, one entry per relative rank:
   `{"rank": "small"|"standard"|"deep", "context_tokens": N, "output_tokens": N}`.
@@ -74,13 +78,13 @@ context to your real model instead of a conservative 32k floor. Report:
   value.
 
 ```bash
-remediate-code next-step --input <path> --host-max-concurrent 4 --host-models '[{"rank":"standard","context_tokens":200000,"output_tokens":32000},{"rank":"deep","context_tokens":200000,"output_tokens":64000}]'
+remediate-code next-step --input <path> --host-can-dispatch-subagents --host-models '[{"rank":"standard","context_tokens":200000,"output_tokens":32000},{"rank":"deep","context_tokens":200000,"output_tokens":64000}]'
 ```
 
 Or with a single dispatch model:
 
 ```bash
-remediate-code next-step --input <path> --host-max-concurrent 4 --host-context-tokens 200000 --host-output-tokens 32000
+remediate-code next-step --input <path> --host-can-dispatch-subagents --host-context-tokens 200000 --host-output-tokens 32000
 ```
 
 Read the returned JSON only far enough to find `prompt_path`, then read and

package/wrapper/audit-code-wrapper-install-hosts.mjs

@@ -104,7 +104,7 @@ function toRepoRelativePath(root, targetPath) {
   return value.length > 0 ? value : '.';
 }
 
-function buildInstallDirective(relativePromptPath) {
+export function buildInstallDirective(relativePromptPath) {
   return [
     INSTALL_MARKER_START,
     '## /audit-code',