recall-os 0.3.0 → 0.3.1

package/README.md

@@ -46,6 +46,10 @@ them.
   the repo, not trapped in an agent's context window.
 - **A gate, not just docs.** `recall doctor` is deterministic and returns an exit code, so "is this
   work actually finished and consistent?" becomes a check you can run in a hook or CI.
+- **Decisions change safely.** When a decision changes, `recall adr supersede` records it (the old
+  ADR is marked superseded, the new one links back) and Doctor flags any memory still citing the old
+  one — so the trail stays auditable instead of silently contradicted. The generated agent rules
+  even carry the CLI commands inline, so your AI tool uses them itself.
 - **Architecture-neutral by design.** Recall OS records and protects _your_ decisions. It never
   silently picks a framework, database, or pattern for you.
 - **Local-first and private.** No network calls, no telemetry, no AI API calls, no remote templates.
@@ -116,6 +120,7 @@ git config core.hooksPath .recall/hooks
 | `recall adr supersede <old> <new>` | Record a changed decision: mark the old ADR superseded by a new accepted ADR. |
 | `recall module create <name>`      | Scaffold module memory (ownership, boundaries, tests).                        |
 | `recall skill create <name>`       | Generate a portable AI agent skill (Claude + Agent Skills).                   |
+| `recall skill list`                | List the built-in agent skill catalog.                                        |
 | `recall mcp add <server>`          | Generate offline, proposed memory for an MCP server.                          |
 | `recall doctor`                    | Validate memory health, evidence, and drift.                                  |
 

package/dist/cli.js

@@ -2946,6 +2946,20 @@ Required reading:
 
 Repository rules override model preferences. If instructions conflict, stop and report the conflict.
 
+## Recall OS commands
+
+This repository is maintained with the Recall OS CLI. Use these commands yourself as you work \u2014 do not
+ask the human to run them, and do not search the web for them (this is a project-local tool):
+
+- \`recall doctor\` \u2014 validate repository memory; run it before claiming any work is complete.
+- \`recall feature create <name>\` \u2014 scaffold feature memory before non-trivial feature work.
+- \`recall adr create <title>\` \u2014 propose a decision; \`recall adr accept <name>\` accepts it.
+- \`recall adr supersede <old> <new-title>\` \u2014 record a changed decision (never overwrite an accepted ADR).
+- \`recall module create <name>\` \u2014 scaffold module memory for a new responsibility boundary.
+- \`recall mcp add <server>\` \u2014 capture an MCP tool's context into memory, offline.
+
+Full command reference: \`docs/ai/RECALL_COMMANDS.md\`.
+
 ## Changing an accepted decision
 
 Before changing anything an accepted ADR governs (framework, database, auth, API shape, and similar):
@@ -2999,6 +3013,11 @@ Before non-trivial work:
 Source-of-truth order: accepted ADRs and repository decisions, then architecture docs, engineering
 standards, the current PRD, security and testing docs, module docs, feature plans, then chat history.
 
+Recall OS commands \u2014 use these yourself (do not web-search this project-local CLI): \`recall doctor\`,
+\`recall feature create <name>\`, \`recall adr create <title>\` then \`recall adr accept <name>\`,
+\`recall adr supersede <old> <new-title>\`, \`recall module create <name>\`, \`recall mcp add <server>\`.
+Full reference: \`docs/ai/RECALL_COMMANDS.md\`.
+
 Before claiming work is complete, run \`recall doctor\` and fix reported errors.
 `
   },
@@ -3534,7 +3553,7 @@ function renderSessionStartHook() {
 adrs=$(ls docs/adrs/ADR-*.md 2>/dev/null | sed 's|.*/||;s|\\.md$||' | tr '\\n' ' ')
 modules=$(ls -d docs/30-modules/*/ 2>/dev/null | sed 's|docs/30-modules/||;s|/$||' | tr '\\n' ' ')
 
-context="Recall OS repository memory is the source of truth over chat history. Before non-trivial work, read AGENTS.md and the docs it routes to; repository rules override model preference. Accepted ADRs (docs/adrs/): \${adrs:-none yet}. Modules (docs/30-modules/): \${modules:-none yet}. Run 'recall doctor' before claiming work complete."
+context="Recall OS repository memory is the source of truth over chat history. Before non-trivial work, read AGENTS.md and the docs it routes to; repository rules override model preference. Accepted ADRs (docs/adrs/): \${adrs:-none yet}. Modules (docs/30-modules/): \${modules:-none yet}. Use the Recall OS CLI commands listed in AGENTS.md (recall feature/adr/module create, recall adr accept and supersede, recall doctor) yourself; do not web-search them. Run 'recall doctor' before claiming work complete."
 
 printf '{"hookSpecificOutput":{"hookEventName":"SessionStart","additionalContext":"%s"}}\\n' "$context"
 `;