loki-mode 11.74.1 → 11.75.1

package/README.md

@@ -138,7 +138,7 @@ loki control backfill .  # load existing ./.loki/runs
 loki control status      # reachable? how many runs?
 ```
 
-While it is running, runs on this machine ship to it automatically. Guides: [docs/v10/CONTROL-PLANE.md](docs/v10/CONTROL-PLANE.md), [container, Helm and ECS](docs/control-plane-container.md), [migrating from the old dashboard](docs/control-plane-migration.md).
+While it is running, runs on this machine ship to it automatically. Guides: [container, Helm and ECS](docs/control-plane-container.md), [migrating from the old dashboard](docs/control-plane-migration.md).
 
 ## Requirements
 

package/SKILL.md

@@ -3,7 +3,7 @@ name: loki-mode
 description: Runs an autonomous software factory that knows what it is supposed to deliver and proves it did, with a signed receipt. Use when the user says Loki Mode or asks to build, fix or verify software autonomously.
 ---
 
-# Loki Mode v11.74.1
+# Loki Mode v11.75.1
 
 **Current entry points (Loki 10 engine):** `loki "<task>"`, `loki owner/repo#N` and `loki quick "<task>"` run the Loki 10 engine and end with a signed receipt (`loki verify`). `loki backlog owner/repo --all` runs many issues in parallel. Guide: `docs/v10/GUIDE.md`. The skill instructions below, and `loki start`, describe the previous bash loop. PRD-file starts (`loki start ./prd.md`) and the opencode provider still run on it until FC38-SWEEP ports them; the legacy-engine command and the LOKI_ENGINE switch remain removed. Prefer `loki owner/repo#N` for issues.
 
@@ -208,7 +208,6 @@ Pick modules from this table, one hop from here. Each module over 100 lines open
 | Legacy healing, modernization | [healing.md](skills/healing.md) |
 | Plan deepening, knowledge extraction | [compound-learning.md](skills/compound-learning.md) |
 | Managed Agents memory, council, flag hierarchy | [memory.md](skills/memory.md) |
-| Non-trivial change (>3 files, MINOR/MAJOR) | [sdlc-fleet.md](skills/sdlc-fleet.md) |
 | Several agent teams at once, WIP limits | [factory-operations.md](skills/factory-operations.md) |
 | Release, version bump, publish check | [release-cadence.md](skills/release-cadence.md) |
 | Adding your own gate, reviewer or agent | [extending.md](skills/extending.md) |
@@ -476,4 +475,4 @@ Implemented, planned and deprecated items, with the version each landed in: `ski
 
 ---
 
-**v11.74.1 | [Autonomi](https://www.autonomi.dev/) flagship product | core skill**
+**v11.75.1 | [Autonomi](https://www.autonomi.dev/) flagship product | core skill**

package/VERSION

@@ -1 +1 @@
-11.74.1
+11.75.1

package/autonomy/completion-council.sh

@@ -19,7 +19,7 @@
 #   - Multi-agent debate: Voting beats unanimous (+13.2%), KS adaptive stopping
 #   - NVIDIA ToolOrchestra: Efficiency metrics for agent tool use
 #
-# Inline Python (D7): the council runs with the cwd inside the agent's repo, so
+# Inline Python: the council runs with the cwd inside the agent's repo, so
 # every "python3 -c" / "python3 -" here runs -E (a PYTHONPATH empty component
 # would load a committed sitecustomize.py) and drops '' and '.' from sys.path
 # before any other import (a committed json.py would print the verdict).
@@ -251,7 +251,7 @@ council_augment_from_managed_memory() {
         return 0
     fi
     local target_dir="${TARGET_DIR:-.}"
-    # D7: never fall back to the cwd. The council runs with cwd inside the
+    # never fall back to the cwd. The council runs with cwd inside the
     # agent's own repo, which can ship its own memory/managed_memory package
     # that shadows the real one -- `${PROJECT_DIR:-$(pwd)}` would `cd` into
     # that repo and `python3 -m memory.managed_memory.retrieve` would import
@@ -1412,7 +1412,7 @@ council_checklist_gate() {
     # than the old `|| echo PASS`, which cleared the first hard gate on a broken
     # host. (Absent python is a real deployment problem; failing open here is
     # exactly the fake-green this gate exists to prevent.)
-    # S-141: resolved -I -S interpreter (see _loki_snapshot_py_tool); a bare
+    # resolved -I -S interpreter (see _loki_snapshot_py_tool); a bare
     # python3 -E still runs user-site .pth lines that can forge json.load.
     local _gate_py
     if ! _gate_py="$(_loki_snapshot_py_tool)"; then
@@ -1563,7 +1563,7 @@ council_heldout_gate() {
     # The failing titles are NOT carried in this line (a checklist title may
     # contain ':' or '|'); they are read separately from the held-out JSON block
     # below in the BLOCK branch.
-    # S-141: resolved -I -S interpreter. None resolvable -> "BLOCK 0 1" and fall
+    # resolved -I -S interpreter. None resolvable -> "BLOCK 0 1" and fall
     # through, so the BLOCK branch still records its trust event and block file.
     local gate_result _heldout_py
     if ! _heldout_py="$(_loki_snapshot_py_tool)"; then
@@ -1890,7 +1890,7 @@ _council_trivial_diff() {
 }
 
 council_evidence_gate() {
-    # P2 (S-116): when the caller sets _LOKI_EVIDENCE_REQUIRE_TESTS=1, inconclusive
+    # P2: when the caller sets _LOKI_EVIDENCE_REQUIRE_TESTS=1, inconclusive
     # test evidence blocks instead of passing through. council_evaluate sets it as
     # a function-local (bash dynamic scope, gone on return), so the vote alone can
     # never approve while every other caller keeps the old pass-through rc.
@@ -4283,7 +4283,7 @@ council_evaluate() {
         if declare -f loki_council_dispatch_agents >/dev/null 2>&1; then
             if loki_council_dispatch_agents "$ITERATION_COUNT" "${COUNCIL_PRD_PATH:-}"; then
                 local _va_round_file="$COUNCIL_STATE_DIR/votes/round-${ITERATION_COUNT}.json"
-                # S-141: resolved -I -S interpreter. Dispatch succeeded, so an
+                # resolved -I -S interpreter. Dispatch succeeded, so an
                 # unreadable round file is CONTINUE, never a fall-through to the
                 # heuristic path that could still reach COMPLETE.
                 local _va_py=""
@@ -4405,7 +4405,7 @@ council_managed_should_stop() {
         return 1
     fi
 
-    # S-196: every reader and the session heredoc run on the resolved -I -S
+    # every reader and the session heredoc run on the resolved -I -S
     # interpreter (see _loki_snapshot_py_tool); -E still loads user-site .pth
     # files. None resolvable -> never start the managed session: Bash fallback.
     # Deliberate consequence: -S hides site-packages, so the anthropic SDK is
@@ -4426,7 +4426,7 @@ council_managed_should_stop() {
 
     # Build session context: diff_summary, test_summary, pending_tasks.
     # Kept deliberately small so we never choke the session budget.
-    # S-196: diff the target project (same root as loki_dir and
+    # diff the target project (same root as loki_dir and
     # LOKI_TARGET_DIR), never PROJECT_DIR, which is Loki's install tree.
     local diff_summary=""
     diff_summary=$(cd "${TARGET_DIR:-.}" 2>/dev/null && git diff --stat 2>/dev/null | tail -20 | tr '\n' ' ' || echo "")
@@ -4472,7 +4472,7 @@ import sys; sys.path[:] = [p for p in sys.path if p not in ("", ".")]
 import json, os, sys, pathlib
 
 # Path setup: import providers.managed only from an explicit PROJECT_DIR. Never
-# fall back to the cwd (D7): it is the agent's repo, which could ship its own
+# fall back to the cwd: it is the agent's repo, which could ship its own
 # providers/managed.py. Unset -> ImportError -> the Bash voting path.
 project_dir = os.environ.get("PROJECT_DIR", "")
 if project_dir:
@@ -4859,7 +4859,7 @@ council_should_stop() {
         # managed memory store. Backgrounded + silent; flags gate the work
         # inside the Python module so no-op when off.
         #
-        # D7: never fall back to the cwd. Identical threat model to
+        # never fall back to the cwd. Identical threat model to
         # council_augment_from_managed_memory (BACKLOG 63) but more severe:
         # this call EXECUTES memory.managed_memory.shadow_write as a module,
         # not just imports one for reading, so `cd "${PROJECT_DIR:-$(pwd)}"`

package/autonomy/council-v2.sh

@@ -19,7 +19,7 @@
 #   - swarm/sycophancy.py (sycophancy detection)
 #   - swarm/calibration.py (reviewer calibration)
 #
-# Inline Python (D7): the council runs with the cwd inside the agent's repo, so
+# Inline Python: the council runs with the cwd inside the agent's repo, so
 # every python here drops '' and '.' from sys.path before any other import (the
 # swarm dir is inserted after that); a committed json.py or sitecustomize.py
 # cannot read the votes. Pinned by tests/test-council-no-fabricated-verdict.sh
@@ -76,7 +76,7 @@ council_v2_vote() {
     local iteration="${4:-0}"
 
     local council_size="${COUNCIL_SIZE:-3}"
-    # S-201: none resolvable -> every read below takes its fail-safe fallback.
+    # none resolvable -> every read below takes its fail-safe fallback.
     local _c2_py
     _c2_py="$(_loki_snapshot_py_tool)" || _c2_py=""
 
@@ -180,7 +180,7 @@ import json
 votes = json.loads(sys.argv[1])
 print('{:.3f}'.format(detect_sycophancy(votes)))
 " "$votes_json" 2>/dev/null) || sycophancy_score=""
-    # S-201: a crashed or missing detector did not measure 0.000. Record the
+    # a crashed or missing detector did not measure 0.000. Record the
     # score as unmeasured (null keeps summary.json valid JSON) so Step 5
     # challenges instead of reading "independent".
     case "$sycophancy_score" in
@@ -194,7 +194,7 @@ print('{:.3f}'.format(detect_sycophancy(votes)))
     if [ "$approve_count" -eq "$council_size" ]; then
         local threshold="${LOKI_COUNCIL_SYCOPHANCY_THRESHOLD:-0.6}"
         local should_challenge
-        # S-201: an unmeasured score, or a compare that cannot run, challenges.
+        # an unmeasured score, or a compare that cannot run, challenges.
         # Only a measured score below the threshold skips the devil's advocate.
         if [ "$sycophancy_score" = "null" ]; then
             should_challenge="yes"

package/autonomy/grill.sh

@@ -248,7 +248,7 @@ grill_invoke_provider() {
             _LOKI_PERM_ARGV=(--dangerously-skip-permissions)
             ! type _loki_claude_permission_argv >/dev/null 2>&1 || _loki_claude_permission_argv
             local _gr_argv=("${_LOKI_PERM_ARGV[@]}" "--model" "${LOKI_GRILL_MODEL:-sonnet}")
-            # S8: lean child flags (text-only role, no tools/MCP/slash commands) and
+            # lean child flags (text-only role, no tools/MCP/slash commands) and
             # a ledger line per call via --output-format json (unwrapped below).
             local _gr_lf _gr_ledger=false
             if type loki_lean_child_flags >/dev/null 2>&1; then

package/autonomy/lib/browser-open.sh

@@ -1,5 +1,5 @@
 #!/usr/bin/env bash
-# Loki Mode -- the single guarded browser opener (S-103).
+# Loki Mode -- the single guarded browser opener.
 #
 # Every place Loki opens a browser MUST go through loki_open_url. Tests and
 # agent sessions must never pop a browser tab on a developer's machine (a

package/autonomy/lib/chat.sh

@@ -72,7 +72,7 @@ HLP
 }
 
 
-# /steer: record source=chat and the chat user as actor (D02). Echo loki steer's own
+# /steer: record source=chat and the chat user as actor. Echo loki steer's own
 # "Steer rev N received" ack; never claim receipt it did not print.
 _chat_steer() {
     local loki="$1" msg="$2" out rc=0 first

package/autonomy/lib/cockpit-render.sh

@@ -239,7 +239,7 @@ cockpit_render() {
 }
 
 #===============================================================================
-# E3: keyboard interactivity for `loki cockpit --follow`.
+# keyboard interactivity for `loki cockpit --follow`.
 # These live here (sourced by cmd_cockpit) so they can be extracted + unit
 # tested without pulling in the whole autonomy/loki CLI.
 #===============================================================================

package/autonomy/lib/done-recognition.sh

@@ -37,7 +37,7 @@
 #
 # No emojis. No em dashes. bash 3.2 safe. Honors `set -uo pipefail`.
 #
-# Inline Python (D7): the gate runs with the cwd inside the agent's repo, so
+# Inline Python: the gate runs with the cwd inside the agent's repo, so
 # every python3 here drops '' and '.' from sys.path before any other import; a
 # committed json.py or sitecustomize.py cannot turn the verdict done. Pinned by
 # tests/test-reuse-done-recognition.sh case (d7).
@@ -318,7 +318,7 @@ reuse_done_recognition_gate() {
     fi
     local _test_results="$loki_dir/quality/test-results.json"
 
-    # --- Resolve the -I -S interpreter (S-200); none -> inconclusive -> build.
+    # --- Resolve the -I -S interpreter; none -> inconclusive -> build.
     local _dr_py
     if ! _dr_py="$(_loki_snapshot_py_tool)"; then
         log_info "Done-recognition: no isolated python3 interpreter resolved; proceeding to build."
@@ -776,7 +776,7 @@ _loki_done_recog_finish() {
     local parsed="$2"
     local loki_dir="${TARGET_DIR:-.}/.loki"
 
-    # S-200: no -I -S interpreter -> write no done record (the gate then builds).
+    # no -I -S interpreter -> write no done record (the gate then builds).
     local _dr_py
     _dr_py="$(_loki_snapshot_py_tool)" || return 1
 
@@ -923,7 +923,7 @@ _loki_done_recog_write_manifest() {
     local parsed="$2"
     local loki_dir="${TARGET_DIR:-.}/.loki"
 
-    # S-200: no -I -S interpreter -> record nothing as met (full build).
+    # no -I -S interpreter -> record nothing as met (full build).
     local _dr_py
     _dr_py="$(_loki_snapshot_py_tool)" || return 1
 

package/autonomy/lib/efficiency_cost.py

@@ -217,7 +217,7 @@ def collect_efficiency(loki_dir):
     # real observation, and nulling it would be its own dishonesty (existing
     # contract: test_genuine_zero_cost_stays_zero_not_null). What is unknown is a
     # record that never carried the field at all.
-    # FC-44: a record flagged tokens_measured:false (an ambiguous resumed session) carries no usage keys. One of them makes every
+    # a record flagged tokens_measured:false (an ambiguous resumed session) carries no usage keys. One of them makes every
     # token total NOT RECORDED (None); the flag keys are added only then, so a complete run's dict is unchanged.
     _unmeasured_tokens = sum(1 for r in _records if r.get("tokens_measured") is False)
     _any_cost = any("cost_usd" in r for r in _records)

package/autonomy/lib/engine10-push.sh

@@ -58,7 +58,7 @@ _LOKI_PINNED_ORIGIN="$_e10_origin"
 
 [ "$_LOKI_ORIGIN_PINNED" = "1" ] || die "origin not pinned (_LOKI_ORIGIN_PINNED=1 required)"
 
-# E-41: the eval harness (eval/loki10/harness.py) pins origin to a local bare
+# the eval harness (eval/loki10/harness.py) pins origin to a local bare
 # repo and scores "PR opened" as a branch landing there. Only an absolute,
 # colon-free path whose own git dir is itself a bare repo qualifies; anything
 # else (URL, scp host:path, relative, non-bare, a subdirectory) takes the
@@ -270,7 +270,7 @@ case "$mode" in
         _e10_gh pr comment "$1" --repo "$_e10_repo" --body-file "$2" || die "gh pr comment failed"
         ;;
     issue-comment)
-        # E-67: a FAILED run with no diff (or no remote) has no PR to open; comment on the issue
+        # a FAILED run with no diff (or no remote) has no PR to open; comment on the issue
         # instead so the run's reason is never silently lost. <issue-ref> is owner/repo#N or #N.
         # --update: the status projector. One comment carrying the marker below is edited in place (created when absent),
         # so an issue shows live run state without a new comment per event. Only a comment written by the authenticated

package/autonomy/lib/prd-enrich.sh

@@ -88,7 +88,7 @@ _loki_prd_enrich_invoke() {
     local out=""
     _LOKI_PERM_ARGV=(--dangerously-skip-permissions)
     ! type _loki_claude_permission_argv >/dev/null 2>&1 || _loki_claude_permission_argv
-    # S8: lean child flags (no user-scope settings/MCP/tools) + ledger line.
+    # lean child flags (no user-scope settings/MCP/tools) + ledger line.
     type loki_lean_child_flags >/dev/null 2>&1 \
         || . "$(dirname "${BASH_SOURCE[0]}")/claude-flags.sh" 2>/dev/null || true
     local _pe_argv=("${_LOKI_PERM_ARGV[@]}") _pe_lf _pe_ledger=false

package/autonomy/lib/proof-check.sh

@@ -27,7 +27,7 @@
 # This function is only ever called when the operator opted in (the call site
 # guards on LOKI_PROVEN_PR_CHECK=1). It is also safe if called directly.
 #
-# Inline Python (D7): the call site's cwd is the agent's repo, so the proof
+# Inline Python: the call site's cwd is the agent's repo, so the proof
 # readers drop '' and '.' from sys.path first; a committed json.py or
 # sitecustomize.py cannot change the headline. Pinned by
 # tests/test-proven-pr-check.sh (D7 section).

package/autonomy/lib/proof-generator.py

@@ -1789,7 +1789,7 @@ def _build_proof(args, loki_dir, target_dir, repo_root):
             "post_headline": True,
         })
 
-    # D47: edited assertion lines are disclosed but never move the headline or rc.
+    # edited assertion lines are disclosed but never move the headline or rc.
     _ae = (facts.get("tests_integrity") or {}).get("assertions_edited")
     if _ae:
         degraded.append({
@@ -1966,7 +1966,7 @@ def _compute_degraded(facts):
         if rec.get("state") == "gap":
             out.append({"item": "functionality:%s" % axis, "status": "failed",
                         "reason": rec.get("reason") or "axis disproven"})
-    # TEN-GATES-10 (D105): the agent writes both the code and the suite, so a
+    # TEN-GATES-10: the agent writes both the code and the suite, so a
     # green suite alone is the agent grading its own work. When the suite is the
     # ONLY passing fact, record the gap here (pre-headline, so proof-verify.py
     # re-derives the same amber headline from the recorded ledger) and VERIFIED
@@ -2102,7 +2102,7 @@ def _compute_headline(facts, degraded):
         and tests.get("exit_code") == 0
     )
     if tests_verified and not degraded and diff_nonempty:
-        # S-113: ablation showed deleting the change leaves the suite green, so
+        # ablation showed deleting the change leaves the suite green, so
         # the passing tests do not prove the change. Never VERIFIED; amber.
         ablation = facts.get("ablation")
         if isinstance(ablation, dict) and ablation.get("status") == "not_load_bearing":

package/autonomy/lib/proof-pr.sh

@@ -32,7 +32,7 @@
 # every optional tool is `command -v`-guarded. All print paths `return 0` so a
 # sourced call cannot abort the caller under set -e.
 
-# S-215: callers render from the agent's repo, and a bare `python3 -` puts the
+# callers render from the agent's repo, and a bare `python3 -` puts the
 # cwd first on sys.path, so a committed json.py forged the headline. The
 # renderer resolves its interpreter through _loki_snapshot_py_tool and runs it
 # -I -S. The copy below only serves callers that source this file on its own;

package/autonomy/lib/skill-link-heal.sh

@@ -1,5 +1,5 @@
 #!/usr/bin/env bash
-# FC-30: keep the provider skill links Loki created pointing at the running
+# keep the provider skill links Loki created pointing at the running
 # install. One shared mechanism, called from npm postinstall and from the bin/loki
 # shim (first invocation after an install or upgrade). Cost: a few lstat/readlink
 # calls, no network, no subprocess beyond readlink/ln/mv on the repair path.

package/autonomy/lib/voter-agents.sh

@@ -27,7 +27,7 @@
 #                                             condition (caller falls through
 #                                             to existing heuristic dispatch).
 #
-# Inline Python (D7): the dispatch runs with the cwd inside the agent's repo, so
+# Inline Python: the dispatch runs with the cwd inside the agent's repo, so
 # every python3 here drops '' and '.' from sys.path before any other import; a
 # committed json.py or sitecustomize.py cannot read the votes.
 # Pinned by tests/test-voter-agents-json.sh (Case D).

package/autonomy/loki

@@ -1220,7 +1220,7 @@ PY
 }
 
 # loki_verify_root: the ONE place that decides which tree the receipt verifier
-# hashes (FC-29). The verifier walks and digests the directory it is given, so
+# hashes. The verifier walks and digests the directory it is given, so
 # an ambient cwd (a huge checkout, $HOME, a non-project directory) must never be
 # the default. Order: explicit TARGET_DIR; the parent of an absolute LOKI_DIR;
 # the cwd if it holds .loki; the repo top-level if it holds .loki. Anything else
@@ -2017,7 +2017,6 @@ show_help_all() {
     echo "Cost controls (budget, iteration and wall-clock caps): docs/cost-controls.md"
     echo "Already have a codebase? Start read-only: docs/brownfield.md"
     echo "No egress allowed? What runs disconnected: docs/air-gapped.md"
-    echo "Where this is going: docs/STRATEGY-2026-2028.md"
     echo ""
     echo "Scripting Loki? Exit codes for every command are documented in"
     echo "docs/exit-codes.md. Severity rises with the code, and 'loki start'"
@@ -2079,7 +2078,7 @@ show_help_aliases() {
     echo "See internal/CLI-CONSOLIDATION-DESIGN.md for the full migration plan."
 }
 
-# D51: bare `loki`. Reuse a running dashboard, else start one on a free
+# bare `loki`. Reuse a running dashboard, else start one on a free
 # loopback port, then open /start (or just print the URL when headless).
 loki_ui_free_port() {
     python3 - "$1" <<'PY'
@@ -2162,7 +2161,7 @@ _loki_control_ui() {
             echo "loki: the Control Plane needs bun (https://bun.sh); using the classic dashboard." >&2
             return 1
         fi
-        # FC-07: LOKI_CONTROL=0 is the test off switch (supervisor, run-all-tests, local-ci);
+        # LOKI_CONTROL=0 is the test off switch (supervisor, run-all-tests, local-ci);
         # it must never start a detached `control serve` that writes ~/.loki/control.
         [ "${LOKI_CONTROL:-}" = "0" ] && return 1
         local cli="$SKILL_DIR/loki-ts/dist/loki.js" cdir="${HOME}/.loki/control" spid
@@ -2388,7 +2387,7 @@ detect_arg_type() {
     echo "unknown"
 }
 
-# --- loki start detach-default handoff (S1) --------------------------------
+# --- loki start detach-default handoff --------------------------------
 # On an interactive TTY, `loki start` now detaches the build to the background
 # by default after a rich, reversible handoff card. A non-interactive shell,
 # --bg, --yes/-y, or CI takes the exact pre-existing foreground path (no card,
@@ -2409,7 +2408,7 @@ _loki_start_should_handoff() {
     return 0
 }
 
-# Stare-worthy brand banner (S6). Renders the real Autonomi logo (purple squircle
+# Stare-worthy brand banner. Renders the real Autonomi logo (purple squircle
 # + white "A" + teal dot) as a crafted truecolor block, next to a "LOKI" wordmark,
 # straight to stderr. ZERO new tooling: pure ANSI escapes. Degrades by color depth
 # ($COLORTERM truecolor -> 24-bit; else a clean 256/mono badge). Never required,
@@ -2461,7 +2460,7 @@ _loki_brand_banner() {
     fi
 }
 
-# E4: a 3-line "what Loki will do" preview from REAL signals only. Line 1 ties
+# a 3-line "what Loki will do" preview from REAL signals only. Line 1 ties
 # the detected tier to its phase count (simple=3, standard=6, complex=8 -- the
 # same mapping run.sh uses; auto/unknown stays honest as "auto-detecting"). Line
 # 2 names the spec, or admits "exploring the codebase" when there is none. Line 3
@@ -2525,7 +2524,7 @@ _loki_start_handoff() {
         esac
     fi
 
-    # Stare-worthy brand banner (S6) then the rich handoff card. Both to stderr so
+    # Stare-worthy brand banner then the rich handoff card. Both to stderr so
     # stdout carries only the decision word.
     _loki_brand_banner "$proj"
     {
@@ -4420,7 +4419,7 @@ cmd_start() {
         loki_project_graph_discover "$_pg_target" || true
     fi
 
-    # S1: detach-to-background by default on an interactive TTY. Fires a rich,
+    # detach-to-background by default on an interactive TTY. Fires a rich,
     # reversible handoff card, then arms the existing --bg path so run.sh
     # backgrounds the runner (it writes the log + PID file and prints its own
     # "Running in Background" banner). The HARD gate below keeps every
@@ -6934,7 +6933,7 @@ cmd_status() {
         fi
     fi
 
-    # P12-19 (D04): freshness of the newest Loki 10 run, same wording as the Bun route
+    # P12-19: freshness of the newest Loki 10 run, same wording as the Bun route
     # ("last event Ns ago (seq N)"). Display-only; silent when there is no run or no parseable event.
     local _e10_last _e10_ev _e10_seq _e10_ts _e10_epoch
     _e10_last=$(ls -d "$LOKI_DIR"/runs/e10-*/ 2>/dev/null | sort | tail -1) || _e10_last=""
@@ -8127,7 +8126,7 @@ _loki_dashboard_args_need_classic() {
 cmd_dashboard() {
     local subcommand="${1:-}"
 
-    # CP-LEGACY (D57): the Control Plane is the default UI for `dashboard`,
+    # CP-LEGACY: the Control Plane is the default UI for `dashboard`,
     # `dashboard start` and `dashboard open`, with or without flags. The legacy
     # Python dashboard stays reachable only by explicit opt-in:
     # LOKI_LEGACY_DASHBOARD=1 or LOKI_CONTROL_DEFAULT=0. Enterprise bind/TLS
@@ -14509,7 +14508,7 @@ cmd_config_schema() {
     loki_config_generate_schema
 }
 
-# D65: multi-repo workspaces (D51 Phase B). On by default; LOKI_WORKSPACES=0 disables.
+# multi-repo workspaces (D51 Phase B). On by default; LOKI_WORKSPACES=0 disables.
 cmd_workspace() {
     if [ "${LOKI_WORKSPACES:-}" = "0" ]; then
         echo -e "${RED}loki workspace is disabled.${NC} Unset LOKI_WORKSPACES (it is set to 0)." >&2
@@ -14519,7 +14518,7 @@ cmd_workspace() {
     python3 "$_LOKI_SCRIPT_DIR/lib/workspace.py" "$@"
 }
 
-# D51: headless backlog runner. One v10 issue-mode run per open issue, each in
+# headless backlog runner. One v10 issue-mode run per open issue, each in
 # its own worktree and branch, at most N in parallel (loki.yaml concurrency).
 cmd_backlog() {
     case "${1:-}" in
@@ -14549,7 +14548,7 @@ cmd_backlog() {
 # failures. Non-zero exit on any failure.
 cmd_config_validate() {
     local file="${1:-}" _loki_yaml
-    # D51: loki.yaml (UI + headless config) has its own schema; no argument
+    # loki.yaml (UI + headless config) has its own schema; no argument
     # validates the discovered loki.yaml (repo root, then ~/.loki/loki.yaml).
     case "$(basename -- "$file")" in "" | loki.yaml | loki.yml | loki.yaml.example) _loki_yaml=1 ;; *) _loki_yaml=0 ;; esac
     if [ "$_loki_yaml" = 1 ]; then
@@ -19767,7 +19766,7 @@ with open(manifest_path, 'w') as f:
 # usable.
 # ---------------------------------------------------------------------------
 cmd_verify() {
-    # D48: --pubkey FILE or --pubkey=FILE (any position) verifies a receipt against a given public key (v10 engine, no ~/.loki needed).
+    # --pubkey FILE or --pubkey=FILE (any position) verifies a receipt against a given public key (v10 engine, no ~/.loki needed).
     local _pk_arg
     for _pk_arg in "$@"; do
         case "$_pk_arg" in
@@ -21326,15 +21325,9 @@ print(json.loads(p.read_text()).get('port', 7373) if p.exists() else 7373)
             python3 "$trigger_schedule_script" remove "$name"
             ;;
         daemon)
-            local daemon_flag=""
-            [[ "${1:-}" == "--dry-run" ]] && daemon_flag="--dry-run"
-            [[ "${1:-}" == "--loop" ]] && daemon_flag="--loop"
-            python3 "$trigger_schedule_script" daemon $daemon_flag
-            ;;
-        install|uninstall)
-            # Always-on company seats: LaunchAgent running the daemon loop
-            LOKI_BIN="${LOKI_BIN:-$(readlink -f "${BASH_SOURCE[0]}")}" \
-                python3 "$trigger_schedule_script" "$subcmd" "$@"
+            local dry_run_flag=""
+            [[ "${1:-}" == "--dry-run" ]] && dry_run_flag="--dry-run"
+            python3 "$trigger_schedule_script" daemon $dry_run_flag
             ;;
         test)
             local event_type="${1:-issues}"
@@ -21445,11 +21438,7 @@ for line in sys.stdin:
             echo "  add schedule <name> <cron_expr> <action> [args...]"
             echo "                        Add a cron-based schedule trigger"
             echo "  remove <name>         Remove a schedule trigger"
-            echo "  daemon [--dry-run|--loop]"
-            echo "                        Run one pass of schedule checks (call from cron), or loop once a minute"
-            echo "  install [--no-load]   Seed the ten company-seat schedules and install the always-on"
-            echo "                        LaunchAgent (dev.autonomi.loki-company; macOS)"
-            echo "  uninstall [--no-load] Remove the always-on LaunchAgent"
+            echo "  daemon [--dry-run]    Run one pass of schedule checks (call from cron)"
             echo "  test <event-type> [--repo owner/repo] [--issue NUM]"
             echo "                        Send a test webhook event (event-types: issues, pull_request, workflow_run)"
             echo "  logs [--tail N]       Show recent trigger event log (default: last 50)"
@@ -22826,7 +22815,7 @@ main() {
     fi
 
     if [ $# -eq 0 ] || { [ $# -eq 1 ] && [ "$1" = "--no-open" ]; }; then
-        # D51: bare `loki` starts the local UI. LOKI_LANDING=1 keeps the v7.28
+        # bare `loki` starts the local UI. LOKI_LANDING=1 keeps the v7.28
         # newcomer landing (tests and scripts that parse it set it).
         if [ "${LOKI_LANDING:-0}" = "1" ]; then
             show_landing
@@ -23239,7 +23228,7 @@ main() {
             cmd_verify "$@"
             ;;
         keys)
-            # D48: public-key export lives in the Bun v10 engine; re-enter the shim on that route.
+            # public-key export lives in the Bun v10 engine; re-enter the shim on that route.
             exec "$_LOKI_SCRIPT_DIR/../bin/loki" keys "$@"
             ;;
         spec)
@@ -30763,7 +30752,7 @@ tokens_known = bool(iterations)
 ctx_file = os.path.join(loki_dir, "context", "tracking.json")
 ctx = load_json(ctx_file)
 # context-tracker.py seeds tracking.json with all-zero totals, so the file
-# existing is not a measurement: only real (nonzero) totals count (S-147).
+# existing is not a measurement: only real (nonzero) totals count.
 if isinstance(ctx, dict) and not tokens_known:
     totals = ctx.get("totals")
     if isinstance(totals, dict) and (totals.get("total_input") or totals.get("total_output")
@@ -34059,7 +34048,7 @@ _docs_generate() {
     fi
     target_path="$(cd "$target_path" && pwd)"
 
-    # S6: a small-tier run (caller exports LOKI_DOCS_TIER) does not pay for the
+    # a small-tier run (caller exports LOKI_DOCS_TIER) does not pay for the
     # eight-doc suite unless docs were asked for explicitly.
     case "${LOKI_DOCS_TIER:-}" in
         small|simple)
@@ -35533,7 +35522,7 @@ cmd_canary() {
         return 2
     fi
     if [ "$subcommand" = "verify" ]; then
-        # D7: this is a verify path. -E stops PYTHONPATH (an empty component
+        # this is a verify path. -E stops PYTHONPATH (an empty component
         # would add the cwd as an absolute path) and a committed
         # sitecustomize.py from acting before the tool runs. python3 is
         # resolved to an absolute path first so an empty PATH component (or a
@@ -38913,7 +38902,7 @@ PYEOF
             if command -v jq &>/dev/null; then
                 jq . "$pj"
             else
-                # D7: this is a verify-adjacent read of proof.json. -E stops
+                # this is a verify-adjacent read of proof.json. -E stops
                 # PYTHONPATH/sitecustomize.py from acting before this line, and
                 # python3 is resolved to an absolute path first so an empty
                 # PATH component cannot let a committed python3 shim in the
@@ -39385,7 +39374,7 @@ print(json.dumps(doc, indent=2))
             local skip_confirm=0
             local visibility="--public"
             local hosted=0
-            # D7: every python3 invocation below reads proof.json from inside
+            # every python3 invocation below reads proof.json from inside
             # the checkout being shared. Resolve an absolute path once so an
             # empty PATH component cannot let a committed python3 shim in that
             # checkout answer instead of the real interpreter; each call site
@@ -39783,7 +39772,7 @@ cmd_docker() {
         # the port-in-use check, so the picked port is ignored on reuse). Run in
         # a subshell so its internal `exit` paths (already-running -> exit 0,
         # errors -> exit 1) never abort the blocking docker run.
-        # CP-LEGACY (D57): open the Control Plane, never the classic server,
+        # CP-LEGACY: open the Control Plane, never the classic server,
         # unless LOKI_LEGACY_DASHBOARD=1 opts back in. Same open gate as before.
         if [ "${LOKI_LEGACY_DASHBOARD:-0}" != "1" ]; then
             local _cp_args=()

package/autonomy/prd-checklist.sh

@@ -18,7 +18,7 @@
 #   LOKI_CHECKLIST_TIMEOUT      - Timeout per check in seconds (default: 30)
 #   LOKI_CHECKLIST_ENABLED      - Enable/disable checklist (default: true)
 #
-# Inline Python (D7): verification runs with the cwd inside the agent's repo,
+# Inline Python: verification runs with the cwd inside the agent's repo,
 # so every python3 here ignores PYTHON* env vars (an empty PYTHONPATH component
 # would load a committed sitecustomize.py) and every -c / - body drops '' and
 # '.' from sys.path before any other import (a committed json.py would print

package/autonomy/receipt_jwt.py

@@ -46,7 +46,7 @@ import logging
 import os
 from pathlib import Path
 
-# Default signer file name; one definition, free of key-ish words (E-156).
+# Default signer file name; one definition, free of key-ish words.
 RECEIPT_SIGNER_BASENAME = "receipt-ed25519.pem"
 
 try: