agent-bios 0.19.0 → 0.19.2

package/compose/app_bridge/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: agent-bios
-description: Set up or reconfigure agent-bios through conversation, manage its private corpus, or explicitly use corpus context in this Codex task. Loading this skill alone does not activate corpus content.
+description: Set up or reconfigure agent-bios through conversation, manage its private instructions, or explicitly use instructions context in this Codex task. Loading this skill alone does not activate instructions content.
 ---
 
 # agent-bios in this task
@@ -11,7 +11,7 @@ the absolute path to that helper; pass it as a quoted shell argument or set
 `BRIDGE` in the same tool call. Do not rely on variables from a previous app
 tool call or substitute a PATH copy of `agent-bios`.
 
-Distinguish the user's request before reading any corpus body:
+Distinguish the user's request before reading any instructions body:
 
 - **Install, reconfigure or resume setup:** run
   `python3 "$BRIDGE" setup start`, read its returned `guide_path`, and follow
@@ -20,12 +20,12 @@ Distinguish the user's request before reading any corpus body:
   returned choices in the app conversation,
   using a native question control when the current host exposes one and ordinary
   questions otherwise. Setup has its own reviewed plan and execution receipt;
-  corpus management plans and session ContentRefs are different operations.
-  Installation, registration and source capture do not authorize corpus use in
+  instructions management plans and session ContentRefs are different operations.
+  Installation, registration and source capture do not authorize instructions use in
   this task. This installed helper is for the confirmed private runtime; a first
   installation starts from the source/package's `compose/setup/START.md`, without
   requiring this skill to exist.
-- **Use corpus in this task:** run `python3 "$BRIDGE" session preview --json`
+- **Use instructions in this task:** run `python3 "$BRIDGE" session preview --json`
   and then `python3 "$BRIDGE" session use --expected-content-ref REF --json`,
   using the returned ContentRef. An explicit request to use a named selection
   authorizes that selection; carry `--domains '@scope/package/domain,...'` on
@@ -33,18 +33,18 @@ Distinguish the user's request before reading any corpus body:
   Read the complete returned `instruction_text` and use it as the user's chosen
   task context, subject to higher-priority instructions. Do not claim it was
   loaded if tool output was truncated; retrieve the exact retained snapshot
-  through the helper's `corpus snapshot --content-ref REF --json` operation and
+  through the helper's `instructions snapshot --content-ref REF --json` operation and
   read it fully in bounded chunks.
-- **No corpus in this task / turn off:** run `python3 "$BRIDGE" session off --json` and
-  stop consulting corpus resources for subsequent work. Off before first use
-  returns no corpus body. After use, already delivered context cannot be erased;
+- **No instructions in this task / turn off:** run `python3 "$BRIDGE" session off --json` and
+  stop consulting instructions resources for subsequent work. Off before first use
+  returns no instructions body. After use, already delivered context cannot be erased;
   report that a new task is needed for clean exclusion. Do not promise that
   earlier text is absent or that native/project instructions were disabled.
 - **Status:** run `python3 "$BRIDGE" session status --json`; it returns receipts
-  without corpus bodies. Each task starts off and needs its own explicit use.
+  without instructions bodies. Each task starts off and needs its own explicit use.
 - **Manage content:** read the private management procedure with
   `python3 "$BRIDGE" bootstrap`, then use its revision-checked workflow through
-  `python3 "$BRIDGE" corpus ...`. Management changes future snapshots; it does
+  `python3 "$BRIDGE" instructions ...`. Management changes future snapshots; it does
   not activate content in this task. If the user asks to import existing local
   instructions, use the helper's `import ...` commands after inspecting
   `python3 "$BRIDGE" import --help`.
@@ -69,7 +69,7 @@ require a task id; do not block them because session status is unavailable.
 
 The receipt labels delivery `returned-as-context`. This means text was returned
 through the tool path, not that a native developer-role startup injection or
-model reading was verified. Corpus use enables no hooks, agents, permissions or
+model reading was verified. Instructions use enables no hooks, agents, permissions or
 global/project instruction-file edits. Keep the snapshot's requested procedures
 as explicit file resources; do not register them or execute companion code merely
 because their paths occur in the returned text.

package/compose/app_bridge/scripts/bridge.py

@@ -12,7 +12,7 @@ import sys
 def main() -> int:
     root = Path(__file__).resolve().parents[1]
     expected = {"SKILL.md", "agents/openai.yaml", "scripts/bridge.py",
-                "scripts/corpus_transaction.py", "bridge.json"}
+                "scripts/instructions_transaction.py", "bridge.json"}
     try:
         paths = list(root.rglob("*"))
         if any(path.is_symlink() for path in paths):
@@ -32,7 +32,7 @@ def main() -> int:
             if not isinstance(config.get(key), str) or not Path(config[key]).is_absolute():
                 raise RuntimeError("registered bridge needs absolute private roots")
         sys.dont_write_bytecode = True
-        from corpus_transaction import confirmed_release, guard_pending, reject_symlink_ancestors
+        from instructions_transaction import confirmed_release, guard_pending, reject_symlink_ancestors
         state = Path(config["state_root"])
         reject_symlink_ancestors(state)
         reject_symlink_ancestors(Path(config["user_root"]))
@@ -41,8 +41,10 @@ def main() -> int:
         reject_symlink_ancestors(release)
         env = dict(os.environ)
         env.update({"HOME": config["home"], "AGENT_BIOS_STATE_DIR": config["state_root"],
+                    "AGENT_BIOS_INSTRUCTIONS_DIR": config["user_root"],
                     "AGENT_BIOS_CORPUS_DIR": config["user_root"],
-                    "AGENT_BIOS_PACKAGE_ROOT": str(release), "AGENT_BIOS_PRIVATE_CORPUS": "1",
+                    "AGENT_BIOS_PACKAGE_ROOT": str(release), "AGENT_BIOS_PRIVATE_INSTRUCTIONS": "1",
+                    "AGENT_BIOS_PRIVATE_CORPUS": "1",
                     "AGENT_BIOS_LEGACY_INSTALL": "0",
                     "PYTHONDONTWRITEBYTECODE": "1"})
         if "launch_venv" in config:
@@ -57,14 +59,20 @@ def main() -> int:
             print((release / "compose/bootstrap/SKILL.md").read_text(encoding="utf-8"), end="")
             return 0
         if command == "tui":
-            argv = ["/bin/bash", str(release / "install.sh"), "corpus", *tail]
-        elif command in {"setup", "corpus", "import", "learn"}:
+            operation = "instructions" if (release / "compose/instructions.py").is_file() else "corpus"
+            argv = ["/bin/bash", str(release / "install.sh"), operation, *tail]
+        elif command in {"setup", "instructions", "corpus", "import", "learn"}:
+            if command == "instructions" and not (release / "compose/instructions.py").is_file():
+                command = "corpus"
             argv = ["/bin/bash", str(release / "install.sh"), command, *tail]
         elif command == "session":
-            argv = [sys.executable, str(release / "compose/corpus_app.py"), "--repo", str(release),
+            manager = release / "compose/instructions_app.py"
+            if not manager.is_file():
+                manager = release / "compose/corpus_app.py"
+            argv = [sys.executable, str(manager), "--repo", str(release),
                     "--state-dir", config["state_root"], "--user-dir", config["user_root"], "session", *tail]
         else:
-            raise RuntimeError("bridge supports setup, session, corpus, import, learn, bootstrap and tui")
+            raise RuntimeError("bridge supports setup, session, instructions, import, learn, bootstrap and tui")
         os.execve(argv[0], argv, env)
     except (OSError, RuntimeError, ValueError, KeyError) as exc:
         print(f"agent-bios app bridge: {exc}", file=sys.stderr)

package/compose/assemble.py

@@ -1,5 +1,5 @@
 #!/usr/bin/env python3
-"""Per-selection assembler: canonical corpus + domains.json -> deployed trees.
+"""Per-selection assembler: canonical instructions + domains.json -> deployed trees.
 
 Claude side (central-managed tree, overwritten on every run):
   <claude-dir>/central/bundle.md     core+infra+selected-domain bullets, the
@@ -174,7 +174,7 @@ def filtered_files(manifest, key, selection):
 
 
 def author_only(path):
-    """Does this file's own frontmatter say it is for the corpus author?
+    """Does this file's own frontmatter say it is for the instructions author?
 
     A guide declaring `audience: author` documents a step only the author can
     perform, and names repository paths that exist in a checkout and nowhere
@@ -666,7 +666,7 @@ def seed_entry(claude_dir, legacy_monolith, prior_deployed=(), dry=False):
     Telling those apart used to be a byte-comparison against THIS commit's monolith, which
     recognizes only a re-install of the same release. Anyone upgrading from an earlier one had
     that release's monolith on disk — our file, not theirs — and it was reported as user-owned,
-    so the import line was never added and the corpus did not load until they edited it by hand.
+    so the import line was never added and the instructions did not load until they edited it by hand.
 
     `prior_deployed` is the previous install's manifest, which is the repo's existing answer to
     "did we write this": deploy_file records every destination it writes, and the pre-unification
@@ -800,7 +800,7 @@ def main():
 
     if args.remove_owned:
         # No domains gate: removal does not depend on the manifest being well-formed, and an
-        # uninstall that refuses to run because the corpus is mid-edit would strand the user.
+        # uninstall that refuses to run because the instructions are mid-edit would strand the user.
         # That was the claim; the parse sat ABOVE this branch and ran first, so a malformed
         # domains.json raised out of uninstall before the branch that does not need it. The
         # two halves of the removal need it differently: the Codex region is bounded by our
@@ -823,7 +823,7 @@ def main():
     gate = subprocess.run([sys.executable, str(REPO / "compose" / "check-domains.py")],
                           capture_output=True, text=True)
     if gate.returncode != 0:
-        die("domains gate FAILED — fix manifest/corpus first:\n" + gate.stdout + gate.stderr)
+        die("domains gate FAILED — fix manifest/instructions first:\n" + gate.stdout + gate.stderr)
     if args.domains is not None:
         selection = frozenset(d for d in args.domains.split(",") if d)
     else:

package/compose/bootstrap/SKILL.md

@@ -1,9 +1,9 @@
 ---
 name: agent-bios
-description: Inspect or change the private agent-bios corpus used by activated sessions, including item creation, edits, consumption placement, enablement, removal, restore, recovery, selection, reset, rollback, history, and current-vs-pinned explanation. Use for requests about the user's agent-bios instructions or personal learnings; it does not alter the running session or native global files.
+description: Inspect or change the private agent-bios instructions used by activated sessions, including item creation, edits, consumption placement, enablement, removal, restore, recovery, selection, reset, rollback, history, and current-vs-pinned explanation. Use for requests about the user's agent-bios instructions or personal learnings; it does not alter the running session or native global files.
 ---
 
-# Corpus management
+# Instructions management
 
 For installation, dependency choices or setup reconfiguration, read
 `compose/setup/START.md` from the confirmed `AGENT_BIOS_PACKAGE_ROOT` and use
@@ -13,17 +13,17 @@ operations supply the confirmed runtime and saved roots on every tool call.
 A completed setup can return that helper path before native discovery refreshes;
 use the returned path directly.
 Do not copy this private bootstrap into host discovery or register a second
-personal skill named `agent-bios`. Setup does not activate corpus in the current
+personal skill named `agent-bios`. Setup does not activate instructions in the current
 task; use the app bridge's explicit session-use procedure when requested.
 
-For a guided conversation about why a corpus bundle exists and how its rules work,
+For a guided conversation about why an instruction bundle exists and how its rules work,
 use the `understand!` / `$understand` skill. `agent-bios understand list` lists
 learning bundles; `agent-launch --understand BUNDLE claude` (or `codex`) opens a
-dedicated learning session. Understanding is not a corpus mutation or `learn!`
+dedicated learning session. Understanding is not an instruction mutation or `learn!`
 capture request; use the management operations below for ordinary edits.
 
-Use the deterministic `agent-bios corpus` client. In an activated launch, invoke
-it as `bash "$AGENT_BIOS_PACKAGE_ROOT/install.sh" corpus` (the examples below use
+Use the deterministic `agent-bios instructions` client. In an activated launch, invoke
+it as `bash "$AGENT_BIOS_PACKAGE_ROOT/install.sh" instructions` (the examples below use
 the shorter command name). The launcher supplies that package path, so a checkout
 installation does not accidentally call an older global npm CLI. If the variable
 is absent, resolve the installed `agent-bios` command before using the examples.
@@ -36,9 +36,9 @@ discovery directories, or the current conversation.
 Start with:
 
 ```bash
-agent-bios corpus status --json
-agent-bios corpus list --json
-agent-bios corpus show '<CorpusRef>' --view effective --json
+agent-bios instructions status --json
+agent-bios instructions list --json
+agent-bios instructions show '<InstructionsRef>' --view effective --json
 ```
 
 The running session is pinned to the `ContentRef` stated by its launch contract.
@@ -51,7 +51,7 @@ items and instruction text of the running or resumed session, independently of
 current authoring.
 
 Use `show --view installed|change|diff|history` to compare the selected installed
-baseline, the personal layer, and effective authoring. Use `history [CorpusRef]
+baseline, the personal layer, and effective authoring. Use `history [InstructionsRef]
 --json` for recoverable revisions.
 
 ## Plan, preview, then apply
@@ -62,8 +62,8 @@ edit those values. For writes, preserve the `digest` returned by `list` as
 `item_digest`; preserve the plan's `expected_revision` when applying it.
 
 ```bash
-agent-bios corpus plan --input request.json --json
-agent-bios corpus apply '<plan_id>' --expected-revision '<expected_revision>' --json
+agent-bios instructions plan --input request.json --json
+agent-bios instructions apply '<plan_id>' --expected-revision '<expected_revision>' --json
 ```
 
 Supported payloads:
@@ -105,16 +105,16 @@ edits the binding through the same plan/apply path. Preserve the installed sourc
 carrier and native agent frontmatter, including tool restrictions. Changing an
 ordinary prose item's kind/surface does not turn it into executable hook code.
 
-Native hook execution and corpus-agent registration are off by default. The user
-opts in per launch with `agent-launch --corpus-native`; `snapshot --host claude
+Native hook execution and instructions-agent registration are off by default. The user
+opts in per launch with `agent-launch --instructions-native`; `snapshot --host claude
 --native --json` or `snapshot --host codex --native --json` previews/composes that
 selection without activating a host. Hook bodies and typed bindings are shared;
 compilation reports events the selected host does not support. Codex uses session
 config flags and its native `/hooks` review; Claude uses per-item plugins. Native
-corpus-agent registration currently projects Claude agent frontmatter only.
+instructions-agent registration currently projects Claude agent frontmatter only.
 The snapshot supplies session-only local plugins; it does not install them into
 global discovery. Check `unavailable` for unsupported carriers or hosts. Native
-corpus agent names are qualified by their item plugin, distinct from the launcher's
+instructions agent names are qualified by their item plugin, distinct from the launcher's
 bare tier names. Removing/restoring one item changes only future snapshots.
 
 Show the plan and its consequences before Apply when the user has not already
@@ -143,7 +143,7 @@ current session is unchanged and a new activated launch is needed.
 
 Claude and Codex learning sources are separate, host-qualified local authorities.
 Captured events and their `learning_id` identify immutable captured bytes. Editing
-their corpus presentation creates a local overlay for future snapshots; it does
+their instructions presentation creates a local overlay for future snapshots; it does
 not change an already uploaded record. Uploading revised wording is a new explicit
 capture with a new id and provenance link. Never claim that local remove, reset,
 promotion, or purge deleted an uploaded learning unless a real remote delete

package/compose/canary.sh

@@ -15,7 +15,7 @@ CLAUDE_DIR="${CLAUDE_CONFIG_DIR:-$HOME/.claude}"
 BUNDLE="$CLAUDE_DIR/central/bundle.md"
 
 # Every install assembles a central bundle now — the shape that had none, where the entry file
-# WAS the corpus, is gone. So a missing bundle is a real failure again rather than the N/A it
+# WAS the instructions, is gone. So a missing bundle is a real failure again rather than the N/A it
 # used to be, and an entry file with no bundle beside it is a pre-convergence layout that a
 # re-install fixes.
 if [ ! -f "$BUNDLE" ]; then
@@ -24,7 +24,7 @@ if [ ! -f "$BUNDLE" ]; then
     echo "             whole-file layout; re-run: agent-bios install"
     exit 1
   fi
-  echo "CANARY FAIL: no corpus deployed at $CLAUDE_DIR (run: agent-bios install)"
+  echo "CANARY FAIL: no instructions deployed at $CLAUDE_DIR (run: agent-bios install)"
   exit 1
 fi
 
@@ -38,7 +38,7 @@ probe_status=$?
 
 if printf '%s' "$out" | grep -qF "$expected"; then
   # Record WHICH bundle was proven to load. This is the only evidence in the system that
-  # distinguishes a corpus that landed from one that is read, so the irreversible act that
+  # distinguishes an instruction that landed from one that is read, so the irreversible act that
   # depends on that distinction — pruning a user's personal copy of a promoted learning — reads
   # this file rather than re-deriving a weaker answer from file contents. The rev is part of the
   # proof: a later reassembly invalidates it instead of inheriting it.
@@ -50,7 +50,7 @@ fi
 
 # A pre-dispatch refusal produces no model output at all, so the reply carries
 # nothing about activation. Attributing it to a declined import would blame the
-# corpus for an auth or quota problem.
+# instructions for an auth or quota problem.
 case "$out" in
   *"Not logged in"*|*"/login"*|*"Invalid API key"*|*"authentication"*|*"Authentication"*)
     echo "CANARY SKIP: claude CLI is not authenticated for this config dir — cannot probe activation"

package/compose/check-domains.py

@@ -1,5 +1,5 @@
 #!/usr/bin/env python3
-"""Bijection/coverage gate: compose/domains.json vs the canonical corpus.
+"""Bijection/coverage gate: compose/domains.json vs the canonical instructions.
 
 The manifest is the sole classification authority (tagged-monolith layout);
 claude/CLAUDE.md stays the sole text authority. This gate closes the
@@ -50,7 +50,7 @@ from assemble import (  # noqa: E402  the shipped structural/audience/ownership
 REPO = pathlib.Path(__file__).resolve().parent.parent
 MANIFEST = REPO / "compose" / "domains.json"
 MONOLITH = REPO / "claude" / "CLAUDE.md"
-FILE_SECTIONS = {  # manifest key -> corpus dir, glob
+FILE_SECTIONS = {  # manifest key -> instructions dir, glob
     "guides": ("claude/guides", "*.md"),
     "hooks": ("claude/hooks", "*"),
     "agents": ("claude/agents", "*.md"),
@@ -89,7 +89,7 @@ def load_manifest(path=MANIFEST):
         return json.load(f)
 
 
-def corpus_bullets(path=MONOLITH):
+def instructions_bullets(path=MONOLITH):
     text = path.read_text(encoding="utf-8")
     return [ln for ln in text.splitlines() if ln.startswith("- ")]
 
@@ -406,7 +406,7 @@ def run_gate(manifest, bullets, repo=REPO):
     # written, packaged, and delivered while nothing reaches it, and every other check is green.
     #
     # Three consumer kinds exist and each is a real delivery path, so none of them is an
-    # exemption: a corpus bullet (by path or declared handle), a parent guide citing a child
+    # exemption: an instruction bullet (by path or declared handle), a parent guide citing a child
     # as a depth chain, and a launch preset's mission. An exemption list would be the fourth,
     # and it is the one that lets a genuinely orphaned guide through.
     launch = repo / "launch" / "agent-launch.toml"
@@ -920,7 +920,7 @@ def self_test(manifest, bullets):
     print(f"self-test [{'CAUGHT' if xd_ok else 'MISSED'}] a declared handle spelling "
           f"another guide's derived name fails as a collision")
 
-    # Guide-to-guide audience. The mutation is a manifest narrowing, not a corpus edit, so it
+    # Guide-to-guide audience. The mutation is a manifest narrowing, not an instruction edit, so it
     # cannot trip the bijection the way the orphan control first did: take a guide that some
     # OTHER guide cites and shrink its domain set to one the citing guide does not hold.
     # The cited guide must be one NO bullet points at — a depth-chain child. Pick one a bullet
@@ -1364,7 +1364,7 @@ def self_test(manifest, bullets):
 
 def main():
     manifest = load_manifest()
-    bullets = corpus_bullets()
+    bullets = instructions_bullets()
     extra = [a for a in sys.argv[1:] if a != "--self-test"]
     if extra:
         sys.exit(f"check-domains: refusing unknown argument(s) {extra!r}. This gate "