Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@docs/data-sources/builtin-toolsets/argocd.md`:
- Around line 161-177: Remove the duplicated tool table by deleting the older
"## Capabilities" section and keep a single canonical "## Tools" table; ensure
the new row for argocd_app_manifest_source_revision is merged into that retained
"## Tools" table so there is only one tools list (include argocd_app_list,
argocd_app_get, argocd_app_diff, argocd_app_manifests, argocd_app_resources,
argocd_app_manifest_source_revision, argocd_app_history, argocd_repo_list,
argocd_proj_list, argocd_proj_get, argocd_cluster_list) and remove the redundant
table block.

In `@docs/data-sources/builtin-toolsets/cilium.md`:
- Around line 174-215: The new "## Tools" table duplicates and conflicts with
the existing "## Capabilities" section (e.g., entries like hubble_observe have
differing details), so consolidate into a single canonical catalog: remove the
duplicate "## Tools" table or merge its rows into the existing "## Capabilities"
table, reconcile divergent entries (ensure hubble_observe flow count and other
descriptions match the canonical source), and if you must keep the old list mark
it explicitly as "legacy" and link to the canonical table; target the headings
and tool names such as "## Tools", "## Capabilities", and specific tool rows
like "hubble_observe" and "cilium_status" when making the changes.

In `@docs/data-sources/builtin-toolsets/kubernetes-remediation-mcp.md`:
- Around line 237-245: The new "## Tools" section duplicates the existing "##
Available Tools" content; remove one and consolidate into a single authoritative
section (preferably keep the existing "## Available Tools" heading), merge the
tool list/table and the upstream MCP server link into that single section, and
ensure any descriptive bullets (Validated kubectl, temporary pods, retrieving
MCP config) are present under the retained header so there's only one source of
truth (look for the "## Tools" and "## Available Tools" headings to locate the
duplicated content).

In `@docs/data-sources/builtin-toolsets/mongodb-atlas.md`:
- Around line 116-123: The docs list a non-existent tool name
"atlas_return_event_type_from_project"; update the documented tool name to the
implemented plural form "atlas_return_events_type_from_project" so it matches
the function/implementation (atlas_return_events_type_from_project) and prevents
callers from referencing a wrong identifier; search for any other occurrences of
the singular name in the same doc and replace them with the plural form.

In `@docs/data-sources/builtin-toolsets/servicenow.md`:
- Around line 204-205: The documentation hardcodes the ServiceNow endpoints as
/api/now/v2/... but the integration supports a configurable api_version
(including empty/unversioned); update the descriptions for
servicenow_get_records and servicenow_get_record to be version-agnostic by
replacing the fixed /api/now/v2/ path with a placeholder or wording that
reflects the configured api_version (e.g., /api/now/{api_version}/table/... or
“/api/now/<configured api_version or unversioned>/table/...”), so the docs match
the api_version configuration option.

---

Outside diff comments:
In `@docs/data-sources/builtin-toolsets/grafanaloki.md`:
- Around line 129-142: The Tools vs Capabilities sections are inconsistent: the
Tools table lists grafana_loki_query while the Capabilities table lists
fetch_pod_logs; reconcile them by either consolidating into one unified section
or removing the outdated entry. Update the markdown so only the actual available
tool(s) are listed (e.g., keep grafana_loki_query and remove fetch_pod_logs if
it was replaced, or list both with clear descriptions if both exist), and ensure
the table headers and introductory text reference the same set of identifiers
(grafana_loki_query, fetch_pod_logs) so readers are not confused.

In `@docs/data-sources/builtin-toolsets/grafanatempo.md`:
- Around line 330-383: The file currently duplicates the same list of tools in
both the "Tools" section and the "Capabilities" section (e.g.,
tempo_fetch_traces_comparative_sample, tempo_search_traces_by_query,
tempo_search_traces_by_tags, tempo_query_trace_by_id, tempo_search_tag_names,
tempo_search_tag_values, tempo_query_metrics_instant,
tempo_query_metrics_range); remove the redundant "Capabilities" section (the
block starting with the second table) and keep the "Tools" section as the single
canonical reference, ensuring the remaining "Tools" table contains the
authoritative descriptions and examples reference the tool names accordingly.

In `@docs/data-sources/builtin-toolsets/kafka.md`:
- Around line 468-495: There are two conflicting tool tables under the headers
"## Capabilities" and "## Tools"; remove the older/duplicate table so only a
single canonical tools inventory remains (keep the comprehensive table that
lists functions like list_kafka_clusters, list_kafka_consumers,
describe_consumer_group, list_topics, describe_topic,
find_consumer_groups_by_topic, consume_messages, list_brokers, describe_configs,
cluster_overview) and delete the other section (the one that lists
kafka_list_topics, kafka_describe_topic, kafka_list_consumers,
kafka_describe_consumer, kafka_consumer_lag) so the docs present one clear API
surface under a single header (e.g., "## Tools").

In `@docs/data-sources/builtin-toolsets/newrelic.md`:
- Around line 119-134: Remove the duplicated "Capabilities" section and keep the
canonical "Tools" table as the single reference; delete the "Capabilities"
header and its table that repeats newrelic_execute_nrql_query and
newrelic_list_organization_accounts (and any identical rows) so maintenance is
centralized under the "Tools" section, preserving any unique notes such as the
enable_multi_account remark into the remaining Tools row if needed.

In `@docs/data-sources/builtin-toolsets/prometheus.md`:
- Around line 450-477: The file currently has two near-identical sections headed
"Tools" and "Capabilities" that duplicate the same 8 entries; remove the
redundant "## Capabilities" section (the header and its following table) and
keep the canonical "## Tools" table as the single source of truth, or if you
prefer both, change the "## Capabilities" heading to clarify a different purpose
(e.g., usage guidance or supported operations) and update the content so it
differs from the "Tools" table; locate these by the "Tools" and "Capabilities"
headings in the Prometheus doc and either delete the entire Capabilities table
block or replace it with distinct content.

In `@docs/data-sources/builtin-toolsets/rabbitmq.md`:
- Around line 58-77: The docs show two conflicting tool tables—keep a single
authoritative table and remove the duplicate/conflicting one; consolidate all
unique tools (e.g., get_rabbitmq_cluster_status, get_rabbitmq_node_info,
get_rabbitmq_queue_info, get_rabbitmq_exchange_info, get_rabbitmq_memory_usage,
get_rabbitmq_disk_usage, list_configured_clusters) into that one table and
ensure the Description text for get_rabbitmq_cluster_status matches its detailed
description (split-brain detection, nodes/listeners/partitions) so no entries
are duplicated or contradictory between the Capabilities and Tools sections.

---

Nitpick comments:
In `@docs/data-sources/builtin-toolsets/azure-sql.md`:
- Around line 70-102: The docs have duplicated/overlapping "Capabilities" and
"Tools" sections across aks.md, aks-node-health.md, and azure-sql.md which
confuses readers; pick one approach and make the content consistent: either
remove the "Capabilities" section from each file if "Tools" replaces it, or keep
both but add a short explanatory note under the "Capabilities" heading (or at
top of the file) clarifying that "Capabilities" is a brief capability summary
for quick scanning while "Tools" gives detailed tool-by-tool usage and
configuration; ensure azure-sql.md reconciles the brief vs. detailed tool
descriptions by moving brief summaries into "Capabilities" and keeping full
examples and parameters in "Tools" so each section has distinct, non-duplicative
content.

In `@docs/data-sources/builtin-toolsets/confluence-mcp.md`:
- Around line 249-259: Merge the duplicate "Tools" documentation blocks by
removing the redundant block introduced under the "Tools" header and
consolidating its content into the existing tools/available-tools section (the
earlier block that lists available tools for HolmesGPT); keep a single canonical
"Tools" section that includes the CQL/confluence search, page retrieval (by
ID/title), comment reading, and optional create/update bullets, retain the MCP
server link for authoritative details, and ensure the consolidated section uses
the exact descriptive text from both blocks so no capabilities are lost.

In `@docs/data-sources/builtin-toolsets/github-mcp.md`:
- Around line 557-569: There are two "## Tools" sections describing the same
GitHub MCP toolset; merge them so there's a single source of truth: remove the
duplicate detailed block starting with the second "## Tools" and either
consolidate its content into the original "available tools" section or replace
the duplicate with a short pointer linking to that original section and the
GitHub MCP server docs; ensure the header "## Tools" appears only once and that
the canonical list (the earlier "available tools" documentation) remains the
authoritative description.

In `@docs/data-sources/builtin-toolsets/helm.md`:
- Around line 48-76: The document duplicates the same Helm tool list under two
headings; remove the entire "## Capabilities" section and retain only the
canonical "## Tools" section to avoid future drift—specifically delete the
duplicated table under the "## Capabilities" heading and keep the table under
"## Tools" (ensure headings "## Capabilities" and "## Tools" are the identifiers
you target).

In `@docs/data-sources/builtin-toolsets/internet.md`:
- Around line 37-49: Remove the duplicate "## Capabilities" section and its
table that lists fetch_webpage, keeping only the standardized "## Tools"
section; specifically delete the "## Capabilities" header and its table rows so
the file contains a single authoritative "## Tools" section documenting
fetch_webpage (refer to the "## Capabilities" header, the "## Tools" header, and
the fetch_webpage tool name to locate the duplicated content).

In `@docs/data-sources/builtin-toolsets/notion.md`:
- Around line 80-94: Duplicate entries for the same tool exist under the "##
Capabilities" and "## Tools" sections; remove the duplicate row under the "##
Capabilities" table and keep the canonical entry under "## Tools" describing
fetch_notion_webpage. Locate the "## Capabilities" heading and the table row
that references fetch_notion_webpage and delete that row (or the whole
Capabilities table if it only contains that duplicate), ensuring only the
standardized "## Tools" table contains the fetch_notion_webpage description;
leave the snippet includes (snippets/toolset_capabilities_intro.md) intact if
used elsewhere.

In `@docs/data-sources/builtin-toolsets/robusta.md`:
- Around line 57-89: The file contains duplicated documentation for the same
Robusta tools under the "## Capabilities" heading and again under "## Tools";
remove the redundant "## Capabilities" block (including its table and the
"Multi-Cluster Support" and "External Changes" subsections) and consolidate any
unique, necessary explanatory text into the standardized "## Tools" section so
only the canonical "## Tools" heading remains with entries for
fetch_finding_by_id, fetch_configuration_changes_metadata,
fetch_resource_issues_metadata, and fetch_resource_recommendation.

In `@docs/data-sources/builtin-toolsets/sentry-mcp.md`:
- Around line 248-258: Remove the duplicated "## Tools" section introduced at
the bottom and consolidate its contents into the existing "Available Tools"
table earlier in the doc so there is a single canonical tools reference;
specifically, merge the bullet list items under the "## Tools" heading into the
earlier "Available Tools" table (or replace the table with the bullets if you
prefer that format) and delete the later "## Tools" heading and its bullets to
avoid duplication and drift.

In `@docs/data-sources/builtin-toolsets/slab.md`:
- Around line 45-59: Remove the duplicate "## Capabilities" section and
consolidate the tool entry under a single standardized "## Tools" section:
delete the entire "## Capabilities" block (including its table and the
fetch_slab_document row) and ensure the existing "## Tools" table contains the
canonical fetch_slab_document description; keep only one table with the tool
name fetch_slab_document and its preferred description to avoid redundant
maintenance.

In `@docs/data-sources/builtin-toolsets/splunk-mcp.md`:
- Around line 165-175: There are duplicate "## Tools" sections; remove this
second "## Tools" block and merge its content (the "Common capabilities exposed
to HolmesGPT include:" bullet list) into the original "## Tools" section already
introduced earlier so the page has one consolidated Tools section; update the
earlier "## Tools" content to include any unique bullets from this block and
delete the duplicate header and paragraphs.