manage-feature-flags — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited manage-feature-flags (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Create, list, kill/restore, and delete Harness FME (Split.io-backed) Feature Flags via MCP.
FME flags are workspace-scoped (not project-scoped). You must discover the workspace_id and environment_id before most operations.
Call MCP tool: harness_list
Parameters:
resource_type: "fme_workspace"Call MCP tool: harness_list
Parameters:
resource_type: "fme_environment"
workspace_id: "<workspace_id>"Call MCP tool: harness_list
Parameters:
resource_type: "fme_feature_flag"
workspace_id: "<workspace_id>"Filter by name, tags, or rollout status:
Call MCP tool: harness_list
Parameters:
resource_type: "fme_feature_flag"
workspace_id: "<workspace_id>"
name: "dark_mode"
tags: "ui"
rollout_status_id: "<uuid>" # discover via fme_rollout_statusCall MCP tool: harness_get
Parameters:
resource_type: "fme_feature_flag"
workspace_id: "<workspace_id>"
feature_flag_name: "<flag_name>"For per-environment targeting/state, use fme_feature_flag_definition and pass environment_id.
Call MCP tool: harness_create
Parameters:
resource_type: "fme_feature_flag"
workspace_id: "<workspace_id>"
traffic_type_id: "<traffic_type_id>" # required by FME
body:
name: "dark_mode"
description: "Enable dark mode UI theme"
tags: ["ui", "rollout"]Kill (turn the flag OFF in one environment):
Call MCP tool: harness_execute
Parameters:
resource_type: "fme_feature_flag"
action: "kill"
workspace_id: "<workspace_id>"
feature_flag_name: "<flag_name>"
environment_id: "<environment_id>"Restore (re-enable after a kill):
Call MCP tool: harness_execute
Parameters:
resource_type: "fme_feature_flag"
action: "restore"
workspace_id: "<workspace_id>"
feature_flag_name: "<flag_name>"
environment_id: "<environment_id>"Call MCP tool: harness_execute
Parameters:
resource_type: "fme_feature_flag"
action: "archive" # or "unarchive"
workspace_id: "<workspace_id>"
feature_flag_name: "<flag_name>"Archiving is subject to OPA policy checks (returns 409 on violation). Unarchive returns 409 if dependent objects exist.
Call MCP tool: harness_update
Parameters:
resource_type: "fme_feature_flag"
workspace_id: "<workspace_id>"
feature_flag_name: "<flag_name>"
body:
description: "Updated description"
tags: ["ui", "ga"]Call MCP tool: harness_delete
Parameters:
resource_type: "fme_feature_flag"
workspace_id: "<workspace_id>"
feature_flag_name: "<flag_name>"| Resource Type | Operations | Description |
|---|---|---|
fme_workspace | list | List FME workspaces |
fme_environment | list | List FME environments (per workspace) |
fme_feature_flag | list, get, create, update, delete, execute(kill/restore/archive/unarchive) | Flag metadata at workspace scope |
fme_feature_flag_definition | list, get | Per-environment rollout targeting and state |
fme_rollout_status | list | Discover valid rollout_status_id values |
fme_rule_based_segment | list, get | Rule-based segments |
fme_rule_based_segment_definition | list, get, execute(enable/disable/change_request) | Segment definition per environment |
fme_feature_flag with body: { name: "dark_mode", ... }kill with production environment_idrestore with staging environment_idfme_feature_flag for workspacearchiveworkspace_id and environment_id before kill/restore — wrong environment can affect production.kill (off) and restore (on) per environment.environment_id is required — kills are per-environmentfeature_flag_name is exact (case-sensitive)workspace_idworkspace_idharness_list with resource_type: "fme_feature_flag" and no filters to see all flags in the workspace~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.