topicforge 0.5.5__tar.gz → 0.6.0__tar.gz

{topicforge-0.5.5 → topicforge-0.6.0}/.gitignore

@@ -160,6 +160,8 @@ dump.rdb
 *.db3
 *.mcap
 rosbag2_*/
+# Test fixtures that are meant to be committed
+!tests/fixtures/bags/**/*.db3
 
 
 # -------------------------------------------------------------------------
@@ -187,42 +189,8 @@ personal/
 CLAUDE*.md
 
 
-# -------------------------------------------------------------------------
-# Project-local docs and drafts kept out of the package / out of clients.
-# Allowlist exceptions: the folder README explains the convention, and
-# `*-spec.md` files are committable specs produced by pack-growth streams
-# (e.g. Stream C of a release plan). Everything else under projet-file/
-# stays local: PDFs, personal market briefs, raw strategy notes.
-# -------------------------------------------------------------------------
-/docs/projet-file/**
-!/docs/projet-file/README.md
-!/docs/projet-file/*-spec.md
-# Traction snapshots are versioned: weekly JSON + summary + folder README.
-# This is the historical curve that informs decision gates G1/G2/G3
-# (product-plan section 12). Numeric history must survive across machines.
-!/docs/projet-file/traction/
-!/docs/projet-file/traction/**
-# Launch-post drafts (Reddit, LinkedIn, X). Versioned so the maintainer
-# can diff drafts across releases and keep marketing history auditable.
-# Drafts, not production copy, never the public README.
-!/docs/projet-file/launch-posts/
-!/docs/projet-file/launch-posts/**
-# Audit reports (security, architecture) + audit-followup triage docs.
-# Versioned so audit trails survive across machines and the triage
-# decisions are bisectable per release.
-!/docs/projet-file/*-audit-*.md
-!/docs/projet-file/audit-*.md
-# External reference material (OMG interop reports, third-party specs
-# snapshots) the maintainer wants pinned in git so strategy decisions
-# remain reproducible. Tracked, not part of sdist.
-!/docs/projet-file/references/
-!/docs/projet-file/references/**
-# Archive of superseded strategic artifacts (old audit reports, dated
-# launch-post drafts). Kept in git so the historical context survives
-# but tucked away so the active `projet-file/` folder stays focused on
-# what is currently load-bearing.
-!/docs/projet-file/archive/
-!/docs/projet-file/archive/**
+# Owner's private strategy notes, specs and traction snapshots: local only.
+/docs/projet-file/
 /docs/assets/screencast-raw/
 
 
@@ -235,3 +203,6 @@ pro/
 scripts/integration/**/rti_license.dat
 scripts/integration/**/*.dat
 .venv-demo/
+
+# Owner's traction snapshot script (runs from a local scheduled task): local only.
+/scripts/traction-snapshot.sh

{topicforge-0.5.5 → topicforge-0.6.0}/CHANGELOG.md

@@ -5,234 +5,262 @@ All notable changes to TopicForge are documented in this file.
 The format is based on [Keep a Changelog 1.1.0](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning 2.0.0](https://semver.org/spec/v2.0.0.html).
 
-## [Unreleased]
+## [0.6.0] - 2026-10-05
+
+`sample_messages` (live) rewritten after the OmniSim team's report (count
+ignored, empty results on latched topics, wrong timestamps and shifted columns
+on simulated time).
+
+### Breaking
+
+- `sample_messages` payloads are nested named fields (`payload.header.stamp.sec`,
+  `payload.ranges`) instead of positional `col_0`, `col_1`, ... columns. The
+  verbatim row under `_raw_text` is gone: it is present only when a message could
+  not be parsed.
+- `_truncated_after_columns` and `_truncated_columns` are replaced by
+  `_truncated_fields`, a list of dotted field paths (`["ranges", "intensities"]`)
+  for arrays, strings or bytes cut at `max_array_length`. A cut array no longer
+  carries the CLI's `'...'` element.
+- `nan`, `inf` and `-inf` floats are returned as the strings `"nan"`, `"inf"`,
+  `"-inf"`.
+- `MiddlewareAdapter.sample_messages` returns a `SampleResult` instead of a list
+  of `MessageSample` and takes `timeout_s`. `parse_csv_echo` and `parse_echo_yaml`
+  are removed.
+- `pyyaml>=6` is a runtime dependency.
+- A failing `ros2 topic echo` (non-zero exit with no message, or a CLI that cannot
+  be started) raises an MCP error instead of returning an empty list. `count` 0 now
+  validates the topic in live mode too, as it already did in the mock.
+
+### Added
+
+- `sample_messages` parameter `timeout_s` (1..45, default 10): a wall deadline for
+  the whole call, topic lookup included. The call returns within about `timeout_s`
+  plus 2 s (stopping the CLI, decoding what arrived).
+- `MessageSample.stamp_source` (`header` or `none`) and `received_ns` (wall clock
+  when TopicForge read the message).
+- A short result carries a `note` (`N of M messages within T s`) that tells "no
+  publisher is announced" from "a publisher exists but nothing arrived in time",
+  hints `count` 1 for a latched topic, and gives the exit code and stderr when the
+  CLI exited early.
+- A message over the per-message size cap (`TOPICFORGE_MAX_SAMPLE_BYTES`, 1 MiB) is
+  dropped while it streams and counted in the `note`, so a large `Image` no longer
+  costs tens of seconds of YAML decoding. Decoding uses libyaml when available and
+  stops at the call deadline.
+- `count` above 50 is capped with a note, and the tool schema declares the
+  maximum.
+
+### Fixed
+
+- `sample_messages` honours `count`: it streams `ros2 topic echo` and stops at
+  `count` messages or the deadline, then stops the process tree (SIGINT then
+  SIGKILL on POSIX, `taskkill /F /T` on Windows) (OmniSim D3). The echo is
+  started with explicit `--qos-reliability` and `--qos-durability` taken from the
+  publishers, so a latched (`transient_local`) topic is no longer returned empty
+  because the CLI picked its QoS against a cold daemon.
+- Stopping the CLI kills the whole process tree even when the launcher already
+  exited (POSIX: always signals the process group; Windows: a Job Object), so no
+  orphan `ros2` process keeps running. The CLI is started with UTF-8 output.
+- The mock `sample_messages` matches the live shape: headerless messages have
+  `timestamp_ns` 0 and `stamp_source` `none`.
+- `timestamp_ns` is the top-level `header.stamp` whatever its value, so a
+  simulated clock (seconds since the simulation start) no longer falls back to 0
+  and shifts every following column by two (OmniSim D4).
+
+## [0.5.6] - 2026-10-02
+
+Fixes from a live run of 0.5.3 and 0.5.5 against OmniSim's simulated
+Clearpath Husky (ROS 2 Humble, Fast DDS, Cyclone backend for the DDS tools),
+reported with ground truth by the OmniSim team.
+
+### Added
+
+- `sample_messages` takes `max_array_length` (1..65536, default 128, null for no
+  cut) and `arrays_summary_only`. A cut is listed under `_truncated_after_columns`
+  in the sample and in `note`.
+- Size caps on returned samples: 1 MiB per message and 4 MiB per call
+  (`TOPICFORGE_MAX_SAMPLE_BYTES` sets the per-message cap). Over-cap messages are
+  dropped and `note` says so. `peek_bag_samples` is capped the same way, and its
+  arrays are cut at 4096 elements.
+- `BagTopicStats` gains `first_timestamp_ns`, `last_timestamp_ns`,
+  `frequency_basis` (`topic_span` or `bag_duration`) and `latched`.
+- `TopicInfo.qos_durability`; `qos_reliability` and `qos_durability` are filled
+  by `get_topic_info` from the publishers' QoS (`mixed` when they disagree).
+- `HealthReport.dds_inactive_reason` says why `dds_backend` is `none`.
+- A real Humble rosbag2 bag from the OmniSim team as a test fixture
+  (`tests/fixtures/bags/omnisim_humble/`).
+
+### Changed
+
+- `list_topics` (live) uses one `ros2 topic list -v` call for publisher and
+  subscriber counts instead of one `ros2 topic info` per topic, falling back to
+  the per-topic calls if the output is not recognized. It leaves QoS null.
+- `rosbags>=0.11.3` is required (0.10 reads a bare `.db3` as ROS 1 and returns
+  message definitions and QoS as plain strings).
+- A latched topic gets no rate only when its messages span under 1 second (a
+  start-up burst such as `/tf_static`); a latched topic published over a longer
+  span keeps its rate. `latched` is unchanged.
+- `analyze_bag` reads per-topic times of an `.mcap` bag only up to 200 MiB and
+  5 s; past either it keeps `bag_duration` rates and says so in the new
+  `BagAnalysis.note`.
+- The `policies_checked` entry for History reads "History (risky only, where
+  announced)".
+- `max_array_length` also cuts strings and bytes to that many characters plus
+  `...`; those cells are listed under `_truncated_columns`.
+- `QosProfile.history` is now optional, and a new `history_note` explains why it
+  is missing. DDS discovery does not carry History (the builtin endpoint data has
+  no such member), so TopicForge reports it only for its own endpoints and for
+  Cyclone DDS peers that set something other than the default. For Fast DDS, RTI
+  and unknown-vendor endpoints it is `null`; for a Cyclone peer, KEEP_LAST depth 1
+  is also `null` because the binding fills missing QoS with that default. Clients
+  that assumed `history` is always a string must handle `null`.
+- A discovered endpoint that announced no History keeps its reliability,
+  durability and the other policies; before, the whole `qos` became `null`.
+- `detect_qos_mismatches` judges the KEEP_ALL-vs-KEEP_LAST History risk only
+  where both sides announced History; elsewhere one hint states that discovery
+  does not carry it, instead of a warning on every pair.
+
+### Fixed
+
+- `sample_messages` with `arrays_summary_only` shifted every CSV column after an
+  array: `<sequence type: float, length: 541>` contains a comma and was split in
+  two. It is one cell now.
+- `sample_messages` returned no samples and no explanation when the echo timed
+  out (for example a large message with `max_array_length` null); `note` now says
+  so.
+- `list_topics` reported 0 publishers and 0 subscribers for a topic missing from
+  `ros2 topic list -v`; it now asks `ros2 topic info` for that topic.
+- `peek_bag_samples` converted a whole numpy array to a list before cutting it
+  at 4096 elements; it now converts only the part it keeps.
+- Fast DDS endpoints no longer report a History taken from the binding's
+  defaults: `detect_qos_mismatches` treats it as not announced, like the other
+  vendors. This path has never run against a real Fast DDS bus.
+- The `tests/fixtures/bags` bag is left out of the sdist.
+- `peek_bag_samples` failed on every Humble `.db3` bag with "Bag contains no
+  type definitions". The reader now gets the type definitions of the distro the
+  bag records, or Humble, and `note` says which. `LaserScan.ranges` and other
+  numeric arrays now come back as lists, not a numpy repr string.
+- `analyze_bag` rates were count / whole-bag duration, 0.1 to 0.7 percent off on
+  periodic topics and meaningless on latched ones (`/tf_static` showed 0.06 Hz,
+  `/rosout` 0.37 Hz). They are now `(n - 1) / (last - first)` per topic, read
+  from the bag when it is readable locally, and latched topics are flagged.
+- `get_topic_info` returned `qos_reliability: null` on every topic although the
+  CLI prints Reliability and Durability per endpoint.
+- `peek_dds_samples('/scan')` reported the topic as not discovered while
+  `rt/scan` and `scan` worked; it now resolves the name like the other DDS tools
+  and says which topic matched.
+- The "DDS module is not active" error said to install the Cyclone binding even
+  when it was installed. It now states the actual cause: backend not selected,
+  binding missing, or adapter failed to start. README wording aligned.
+- CSV `...` truncation cells from `ros2 topic echo` no longer count as data
+  columns.
+- `ros2` output is decoded as UTF-8 on Windows instead of cp1252.
+- `detect_qos_mismatches` no longer calls service, action, `rosout`,
+  `parameter_events` or `ros_discovery_info` topics typos of each other (for
+  example `get_parametersRequest` vs `set_parametersRequest`), and no longer
+  reports them as orphans. Only plain `rt/` topics and bare DDS names are compared.
+- The "differs by N edits" hint is now only given between a writer-only name and
+  a reader-only name; a topic with both sides is never suggested as a typo.
 
 ## [0.5.5] - 2026-10-02
 
-Driven by a blind evaluation: agents given only TopicForge's tools had to
-diagnose live DDS buses with planted faults. The first rounds (on 0.5.4) found
-wrong blame on QoS, missed Liveliness and Ownership faults, hand-joined GUIDs
-and empty results read as "healthy". After the changes below, 16 of 16
-scenarios were diagnosed correctly, with no false alarm on a healthy bus.
+Tool outputs were reworked after testing them with LLM agents on the author's
+16 test scenarios (live DDS buses with planted faults).
+
+### Breaking
+
+- `detect_qos_mismatches` returns a `MismatchScan` envelope instead of a bare
+  list. Migrate by reading `["reports"]` where you used the list.
 
 ### Added
 
-- **`list_endpoints`, the 12th MCP tool.** Returns every announced DDS writer
-  and reader as a typed `EndpointInfo` (role, topic, type, owning participant
-  guid, name and vendor, structured QoS, announcement timestamp) plus a
-  `by_topic` roll-up that flags orphans (`no_reader`, `no_writer`). TopicForge's
-  own endpoints are excluded unless `include_observer`, and counted in
-  `excluded_observer_endpoints`. Endpoints of a participant that left are kept
-  (200 entries, 1 hour) and shown as `departed_writers` / `departed_readers`;
-  `include_departed` lists them. Served by Cyclone and the mock. Fast raises a
-  clear "not supported yet" error.
-- **Continuous discovery tracking on Cyclone.** A daemon thread (0.5 s period)
-  is the only code that reads the builtin discovery topics and feeds in-memory
-  caches that every discovery tool reads. Lifecycle no longer moves only when a
-  tool is called: a node restarted three times shows as 3 `lost` and 4
-  `discovered`, dated by DDS. The 2 s warm-up sleep is gone.
-- Timestamps on `participant_events` and `list_participants`: `announced_ns`,
-  `lost_ns`, `lost_time_source`, and on events `time_source` and `observed_ns`.
-  A DDS source timestamp and a local observation are never mixed.
-- `health_check` gains `now_ns`, `observer_started_ns`, `observed_domain_note`
-  and the tracker status (`tracker_running`, `tracker_passes`, `tracker_errors`,
-  `tracker_last_pass_ns`).
-- `QosProfile` gains optional `liveliness_kind`, `liveliness_lease_ns`,
-  `ownership_kind`, `ownership_strength`, `partitions`, `latency_budget_ns`,
-  `destination_order` and `data_representation`, read from Cyclone discovery.
-  A missing Partition policy is reported as `[""]` everywhere.
-- `peek_dds_samples` on `DCPSPublication`, `DCPSSubscription` and
-  `DCPSParticipant` carries structured `role`, `participant_guid`,
-  `participant_name`, `type_id`, `qos`, `announced_ns` and `is_observer`, and
-  sets `timestamp_ns` from the announcement. `_raw_text` is kept only for a
-  sample with nothing structured to read, truncated to 300 characters.
-- `peek_dds_samples` on a builtin topic returns the cached discovery state, not
-  a stream.
-- Topic filters accept `rt/x` and `x` interchangeably and report which form
-  matched. A filter that matches nothing returns the list of known topics.
-- Examples: the generic role nodes take partition, liveliness and ownership
-  options.
+- `list_endpoints`, the twelfth tool: every announced DDS writer and reader with
+  owning participant, structured QoS and a per-topic roll-up that flags orphans
+  (`no_reader`, `no_writer`). Cyclone and mock only.
+- Continuous discovery tracking on Cyclone, so a node restarted three times
+  shows as 3 `lost` and 4 `discovered` events dated by DDS, not by poll time.
+- `announced_ns`, `lost_ns`, `time_source` and related timestamp fields on
+  `participant_events` and `list_participants`.
+- `health_check` reports `now_ns`, tracker status and an observed-domain note.
+- `QosProfile` gains Liveliness, Ownership, Partition, LatencyBudget,
+  DestinationOrder and DataRepresentation.
+- `peek_dds_samples` on builtin discovery topics returns structured endpoint
+  and participant fields instead of a raw repr string.
+- Topic filters accept `rt/x` and `x` interchangeably; a filter that matches
+  nothing returns the known topics.
+- `detect_qos_mismatches` also reports `matched` and `not_matched` pairs, hints
+  (near-miss topic names, path suffixes, type id differences), and the policies
+  it did and did not check.
 
 ### Changed
 
-- **Breaking: `detect_qos_mismatches` returns a `MismatchScan` envelope instead
-  of a bare list.** To migrate, read `["reports"]` where you used the list. The
-  envelope also carries `matched`, `not_matched`, `hints`, `pairs_checked`,
-  `topics_scanned`, `policies_checked`, `policies_unchecked` and
-  `mode_effective`. Why: an empty list was read as a healthy bus although only
-  four policies were checked, and readers and writers in different partitions
-  were blamed on Reliability.
-- Partition is checked first (`*` and `?` wildcards; a wildcard against a
-  wildcard never matches). A pair it separates is a `not_matched` entry with
-  reason `partition`; differing type names give reason `type_name`. The RxO
-  rules are not run on either. A `not_matched` pair lists the policies that
-  would be incompatible if it matched (`latent_incompatible_policies`).
-- Liveliness (kind and lease), LatencyBudget, Ownership (kind equality),
+- `detect_qos_mismatches` checks Partition first (with `*` and `?` wildcards),
+  then type names, then the RxO rules; Liveliness, LatencyBudget, Ownership,
   DestinationOrder and DataRepresentation join Reliability, Durability and
-  Deadline. History stays a labelled `risky` finding. A policy a side did not
-  announce is skipped, never guessed.
-- `MismatchReport` gains, additively, the reader and writer participant guid and
-  name, type names, `details` (requested and offered value, failed rule) and
-  `unchecked`.
-- Hints cover orphan topics whose names differ by at most 2 edits (compared
-  against all topics), path-suffix matches and XTypes type id differences (a
-  note, never a mismatch). Hints are prioritized and say how many were omitted.
-- A late joiner is not a hint (it is normal on most buses): a `MatchedPair` is
-  flagged `late_joiner` when its writer is VOLATILE and the reader, on the same
-  host, appeared more than 1 s later.
-- `reports`, `matched` and `not_matched` are capped at 200 entries each, with
-  `reports_total`, `matched_total`, `not_matched_total` and `truncated`.
-- `topic_metrics` reports a `status`, and on a user topic says it has no data
-  instead of returning zeros. `peek_dds_samples` on a user topic returns count
-  0 and a note. `ParticipantInfo` and endpoints gain `is_observer`, and
-  `vendor_source` says where a vendor came from.
-- Tool descriptions no longer carry internal history, and state the
-  single-domain and EXCLUSIVE-ownership facts.
-- `EndpointInfo.activity` is reserved (always `None`) with an `activity_note`
-  saying liveness is not observed.
+  Deadline.
+- `topic_metrics` and `peek_dds_samples` on a user topic say they have no data
+  instead of returning zeros or a placeholder.
+- Tool descriptions no longer carry internal history.
 
 ### Fixed
 
-- Infinite durations (cyclonedds reports 9223372036854775807) are normalized
-  to `None`; an infinite Deadline used to surface as a 9.2e18 ns deadline.
-- The Cyclone discovery tracker is stopped at interpreter exit.
-- A race between the tracker and a tool call could mark a live participant as
-  lost for good; a failed read could drop a participant's departure; endpoints
-  of a participant that had just left could stay listed as live. The tracker
-  and every tool now share one lock, and taken samples are never discarded.
-- The cyclonedds Python binding is not thread-safe when it converts QoS: two
-  threads in `take()` at once corrupted the heap on Windows. Every binding call
-  now goes through one process-wide lock, and tool handlers no longer call the
-  binding at all.
-- If every tracker pass fails, tools no longer wait 3 s each: warm-up is bounded
-  once. `health_check` reports failed passes and cache evictions.
-- An unreadable QoS duration is treated as unknown (`QosProfile.unknown_policies`),
-  not as infinite, so it cannot produce a false Deadline incompatibility.
-- Topic-name typo hints are bounded (banded edit distance, orphan cap, call
-  budget), so a bus with a thousand topics does not stall the call.
-- Partition matching was checked against a live Cyclone bus: `*` and `?` are
-  wildcards, `[...]` is literal, and two wildcard expressions never match each
-  other. Pinned by tests.
+- Infinite durations are `None` instead of 9223372036854775807.
+- Races between the tracker and tool calls could mark a live participant lost
+  or keep endpoints of a departed one.
+- Concurrent cyclonedds calls could corrupt the heap on Windows; every binding
+  call now takes one lock.
+- A failing tracker no longer adds 3 s to every tool call.
+- An unreadable QoS duration is treated as unknown, not infinite.
+- Typo hints stay bounded on a bus with a thousand topics.
 
 ### Known limits
 
-- A hung writer (alive, lease renewed, no data) is not observable without a
-  data probe. An opt-in probe is planned for 0.5.6.
-- A crash and a clean leave cannot be told apart, and `lost_ns` is an upper
-  bound of the death (the lease expiry after a crash).
-- A participant that cycles faster than the discovery history depth between two
-  tracker passes can be missed.
-- The Fast DDS backend has never run on a bus, and `list_endpoints` is not
+- A hung writer (alive, no data) cannot be observed without a data probe.
+- A crash and a clean leave look the same; `lost_ns` is an upper bound.
+- A participant that cycles faster than the history depth between two tracker
+  passes can be missed.
+- The Fast DDS backend has never run on a bus; `list_endpoints` is not
   supported there.
-- DDS Security is not supported.
-- User-topic payload decoding is disabled, so `topic_metrics` has no data on
-  user topics.
+- DDS Security is not supported, and user-topic payload decoding is disabled.
 
 ## [0.5.4] - 2026-10-02
 
-First run of the DDS code against a live multi-vendor bus (Windows 11, a
-Python / Cyclone DDS participant and a Rust / Dust DDS participant, TopicForge
-driven by a real MCP client). Until now every DDS adapter had only been
-checked statically. The run exposed four defects that together made the DDS
-module non-functional on Cyclone; all are fixed and pinned by tests.
+First run of the DDS code against a live bus (Windows 11, a Python / Cyclone DDS
+participant and a Rust / Dust DDS participant). It exposed four defects that
+left the DDS module non-functional on Cyclone; all are fixed and tested.
 
 ### Fixed
 
-- Role nodes published below their rate (about 7 Hz for 10 Hz on Windows):
-  they now keep an absolute schedule.
-
-- **`list_participants` now reports the participant name and the hostname on
-  Cyclone.** The name comes from the EntityName QoS and the hostname from the
-  `__Hostname` discovery property; both were always null on a real bus.
-- **Participant GUIDs were never read.** cyclonedds 11.0.1 exposes the builtin
-  key as a `uuid.UUID`, which the extractor did not handle, so every
-  participant collapsed onto a single `unknown` entry.
-- **Vendors were never identified.** The builtin participant sample carries no
-  vendor field. The vendor id is now read from the first two bytes of the GUID
-  prefix, as RTPS recommends; implementations that do not follow that
-  convention (Dust DDS, and RTI by default) still report `unknown`, because
-  the Cyclone Python binding does not expose the vendor id from the RTPS
-  header.
-- **The OMG vendor-id table was wrong.** It mapped `01.05` to Fast DDS and
-  `01.16` to Cyclone; the correct ids are `01.0F` (eProsima) and `01.10`
-  (Eclipse), verified against both vendors' sources. The Cyclone side ran on a
-  live bus; the Fast DDS side (`fast_extract_vendor_id`) is verified
-  statically only, since its tests need a binding that is not on PyPI. The `vendor` field of
-  `ParticipantInfo` and `ParticipantEvent` now also accepts `rti_micro`,
-  `opensplice`, `opendds`, `coredx`, `intercom` and `dust` (soft-breaking for
-  clients validating the previous enum).
-- **`detect_qos_mismatches` never reported anything on Cyclone.** cyclonedds
-  scopes its policy class names (`Reliability.BestEffort`), and the
-  normalizer matched only the bare name, so no QoS profile was ever built.
-- **A stopped participant never disappeared.** Discovery readers keep the last
-  sample of a departed participant with a NOT_ALIVE instance state; those are
-  now ignored for participants and endpoints, so a participant is reported as
-  left when its lease expires, and a dead endpoint no longer produces a
-  mismatch.
-
-- Cyclone adapter created a new DDS reader on a builtin discovery topic on
-  every tool call and never deleted it. It now keeps one reader per builtin
-  topic and takes a non-blocking snapshot, so calls no longer wait 2 s each
-  (`detect_qos_mismatches` waited 4 s).
-- `peek_dds_samples` on `DCPSPublication` / `DCPSSubscription` reported the
-  endpoints of participants that had left; disposed entries are now dropped.
-  Its payload also carries the endpoint `type_name`.
-- `test_dds_cross_vendor.py` expected an error message from v0.3; it had
-  never run, since CI has no DDS binding. First run against the real Cyclone
-  binding.
+- `list_participants` reports the participant name and hostname on Cyclone
+  (both were always null).
+- Participant GUIDs were never read, so every participant collapsed onto one
+  `unknown` entry.
+- Vendors were never identified. The vendor id is now read from the GUID prefix;
+  Dust DDS and RTI by default still report `unknown`.
+- The OMG vendor-id table mapped Fast DDS and Cyclone to the wrong ids (now
+  `01.0F` and `01.10`). The Fast DDS side is verified statically only.
+- `ParticipantInfo.vendor` also accepts `rti_micro`, `opensplice`, `opendds`,
+  `coredx`, `intercom` and `dust`.
+- `detect_qos_mismatches` never reported anything on Cyclone because policy
+  class names were not normalized.
+- A stopped participant never disappeared; departed participants and endpoints
+  are now dropped.
+- The Cyclone adapter leaked a reader per tool call and waited 2 s each; it now
+  keeps one reader per builtin topic.
+- `peek_dds_samples` on `DCPSPublication` / `DCPSSubscription` listed endpoints
+  of participants that had left, and now includes the endpoint `type_name`.
+- Example role nodes published below their rate on Windows.
 
 ### Added
 
-- `examples/dds/`: a "write the code" track. Examples 00 (hello publisher
-  and subscriber), 05 (Reliability), 06 (Durability and the late joiner),
-  07 (Deadline declared and kept) and 08 (a crash seen from inside and from
-  outside) each ship a readable `publisher.py` and `subscriber.py` in Cyclone
-  DDS Python; the subscriber prints what it receives and the DDS statuses,
-  and TopicForge explains the same situation from outside. All fourteen
-  examples pass on a live bus.
-- The generic role nodes print one line per second per reader with what
-  they received, and examples 02, 04, 10 and 13 check that the broken
-  subscriber receives nothing while the control one receives data.
-- The harness keeps each program's output in a log file and prints it live
-  under `--hold`.
-
-- `scripts/integration/interop_check.py`: one-command multi-vendor demo
-  that starts a Rust / Dust and a Python / Cyclone participant, drives
-  TopicForge over stdio through the official MCP client, checks participant
-  discovery, a deliberate Reliability mismatch between the two vendors, and
-  the departure of a stopped participant, then stops every process it started.
-- A real Rust / Dust DDS participant (`publishers/dust_publisher`) and a real
-  Python / Cyclone participant replacing the previous scaffold, which never
-  wrote a sample.
-- Twelve interop programs in total, one per vendor and language with an
-  officially released binding (Cyclone C / C++ / Rust / Python, Dust Rust /
-  Python, Fast DDS C++ / Python, RTI Connext C / C++ / Python, OpenSplice C),
-  all following the contract in `scripts/integration/DEMO_CONTRACT.md`. The
-  driver starts whichever ones are built on the host and adapts its checks;
-  `--list` shows what can run. Only Cyclone Python and Dust Rust / Python / C
-  have been run, on Windows; the rest are written but unrun. RTI participants
-  need a local license and are never run in CI.
-- One-command launch scripts, `scripts/integration/launch/setup` and
-  `run_demo` (`.ps1` and `.sh`), which create `.venv-demo`, install the Cyclone
-  binding and build the Rust participant. `setup.ps1 -Firewall` adds inbound
-  UDP 7400-7500 rules on private networks for multi-machine runs.
-- Unicast peer configuration for Cyclone and Fast DDS and a guide for a mixed
-  Linux and Windows bus (`scripts/integration/config/`).
-- `.github/workflows/demo.yml` runs the driver with the Cyclone and Dust
-  participants on Ubuntu and Windows; `demo-fast.yml` builds Fast DDS 3 from
-  pinned tags and starts the C++ participant (weekly and manual).
-- `scripts/integration/README.md` rewritten around the demo: it previously
-  described the removed docker / scenario rig.
-
-- `examples/dds/`: real use cases 10 to 14 (driver swap, wiring, safety
-  monitor dropout, deadline not offered, restart loop) next to the concept
-  examples 01 to 04. All nine pass on a live bus.
+- `examples/dds/`: fourteen runnable examples (concepts and use cases), all
+  passing on a live bus.
+- `scripts/integration/`: a demo driver, twelve interop programs (Cyclone, Dust,
+  Fast DDS, RTI, OpenSplice), one-command setup scripts and CI workflows. Only
+  Cyclone Python and Dust Rust / Python / C have been run, on Windows.
 
 ### Removed
 
-- The docker / scenario integration rig (`scenarios_runner.py`, `run-local.*`,
-  `docker-compose.yml`, per-vendor Dockerfiles, scenario JSON files, schema
-  test and `integration.yml`) is removed in favour of the demo driver.
+- The docker / scenario integration rig, replaced by the demo driver.
 
 ## [0.5.3] - 2026-10-01
 
@@ -1235,7 +1263,9 @@ Initial MVP release of TopicForge: ROS Topic Inspector & Bag Analyzer MCP server
 - The write path (publishing, commanding robots) is intentionally out of scope for the MVP.
 - `analyze_bag` in live mode parses `ros2 bag info` text output; deeper anomaly detection remains mock-only for now.
 
-[Unreleased]: https://github.com/yaniswav/TopicForge/compare/v0.5.5...HEAD
+[Unreleased]: https://github.com/yaniswav/TopicForge/compare/v0.6.0...HEAD
+[0.6.0]: https://github.com/yaniswav/TopicForge/compare/v0.5.6...v0.6.0
+[0.5.6]: https://github.com/yaniswav/TopicForge/compare/v0.5.5...v0.5.6
 [0.5.5]: https://github.com/yaniswav/TopicForge/compare/v0.5.4...v0.5.5
 [0.5.4]: https://github.com/yaniswav/TopicForge/compare/v0.5.3...v0.5.4
 [0.5.3]: https://github.com/yaniswav/TopicForge/compare/v0.5.2...v0.5.3