Skip to content
Generated from guardlink --help. Report problems against guardlink, not this page.

Export and report commands

Turn the threat model into reports, diagrams, SARIF, diffs, flow paths, and an AI bill of materials, query it, ask git who wrote each claim, and join models across repositories and applications.

Generate a threat model report with Mermaid diagram

guardlink report [options] [dir]

Arguments

Argument Description
dir Project directory to scan (default: “.”)

Options

Option Description
-p, --project <n> Project name (default: the name in .guardlink/config.json)
-o, --output <file> Write report to file
-f, --format <fmt> Output format: md, json, or both (default: md) (default: “md”)
--diagram-only Output only the Mermaid diagram, no report wrapper
--json Also output threat-model.json alongside the report (legacy; prefer –format)
--feature <names> Filter report to specific feature(s) (comma-separated)
--blame Add an Attribution section: who introduced and fixed each claim, per person and per AI tool (read from git)
--agents Write only the agent threat model: what each agent and principal can reach, unentitled reaches, ungated mutations and the OWASP LLM mapping (default file threat-model-agents.md)
--code-graph Add a measured-route column for each declared agent reach, read from a code graph when one is installed and built (a claim to check, never evidence)
-h, --help display help for command

Emit .guardlink/model.json and .guardlink/graph/ — diagrams and the model as plain files

guardlink artifacts [options] [dir]

Arguments

Argument Description
dir Project directory to scan (default: “.”)

Options

Option Description
-p, --project <n> Project name (default: the name in .guardlink/config.json)
--dry-run Show what would be written without writing
-h, --help display help for command

Export findings as SARIF 2.1.0 for GitHub Advanced Security, VS Code, etc.

guardlink sarif [options] [dir]

Arguments

Argument Description
dir Project directory to scan (default: “.”)

Options

Option Description
-p, --project <n> Project name (default: the name in .guardlink/config.json)
-o, --output <file> Write SARIF to file (default: stdout)
--min-severity <sev> Only include unmitigated exposures at or above this severity (critical|high|medium|low); @confirmed results are always included
--no-diagnostics Exclude parse errors from SARIF output
--profile <name> github (default): findings for code scanning. pentest: also covered exposures, boundary claims and hypothesis-ledger state, appended after the github results (default: “github”)
--baseline <file> Pentest profile only: an earlier SARIF export; sets each result’s baselineState (new, unchanged, updated)
-h, --help display help for command

Compare threat model against a git ref — find what changed

guardlink diff [options] [ref]

Arguments

Argument Description
ref Git ref to compare against (commit, branch, tag, HEAD1) (default: “HEAD1“)

Options

Option Description
-d, --dir <dir> Project directory (default: “.”)
-p, --project <n> Project name (default: the name in .guardlink/config.json)
--markdown Output as markdown (for PR comments)
--json Output as JSON
--fail-on-new Exit 1 if new unmitigated exposures found (CI mode)
-h, --help display help for command

Find source-to-sink flow paths with no control on them (derived from @flows, no LLM)

guardlink paths [options] [dir]

Arguments

Argument Description
dir Project directory (default: “.”)

Options

Option Description
-p, --project <n> Project name (default: the name in .guardlink/config.json)
--all Include paths a control already defends
--boundary-only Only paths that cross a declared @boundary
--json Output as JSON
--fail-on-found Exit 1 if any path is reported (CI mode)
-h, --help display help for command

Query the threat model with the guardlink_lookup forms, e.g. “unentitled reaches”, “agents”, “effects for #db”, “cwe:CWE-89”. Prints JSON; an unrecognised form lists the supported ones

guardlink lookup [options] <query...>

Arguments

Argument Description
query Query form

Options

Option Description
-d, --dir <dir> Project directory (default: “.”)
-p, --project <n> Project name (default: the name in .guardlink/config.json)
--fail-on-found Exit 1 if the query returns any result (CI mode, e.g. no unentitled reaches)
-h, --help display help for command

Who introduced, declared and fixed each claim, and which AI tool co-authored it — read from git, nothing written

guardlink blame [options] [dir]

Arguments

Argument Description
dir Project directory to scan (default: “.”)

Options

Option Description
-p, --project <n> Project name (default: the name in .guardlink/config.json)
--file <path> Only claims declared in, or anchored to, this file (root-relative)
--json Emit the guardlink.blame/v1 payload instead of text
--identity <mode> How people are shown: name (default), email, or hash (overrides blame.identity in config.json)
-h, --help display help for command

Generate an interactive, self-contained HTML threat model dashboard

guardlink dashboard [options] [dir]

Arguments

Argument Description
dir Project directory to scan (default: “.”)

Options

Option Description
-p, --project <n> Project name (default: the name in .guardlink/config.json)
-o, --output <file> Output file (default: threat-dashboard.html)
--light Pin the light theme (by default the page follows the OS preference)
--feature <names> Filter dashboard to specific feature(s) (comma-separated)
--blame Add an Attribution page: exposures introduced and fixed per person and per AI tool (read from git)
--since <ref> Add a “what changed” strip: exposures added and resolved, and claims gone stale, since a git ref (tag, branch, commit)
--code-graph Add a measured-route table to the Agents page, read from a code graph when one is installed and built (a claim to check, never evidence)
-h, --help display help for command

Export agent configuration and declared reach as a CycloneDX 1.6 inventory

guardlink ai-bom [options] [dir]

Arguments

Argument Description
dir Project directory (default: “.”)

Options

Option Description
-o, --output <file> Write JSON to a file
-h, --help display help for command

Translate GuardLink threats into CERT-X-GEN pentest templates (generation only, no execution)

guardlink translate [options] [prompt...]

Arguments

Argument Description
prompt Optional translation instructions

Options

Option Description
-d, --dir <dir> Project directory (default: “.”)
-p, --project <n> Project name (default: the name in .guardlink/config.json)
--claude-code Launch Claude Code in foreground
--codex Launch Codex CLI in foreground
--gemini Launch Gemini CLI in foreground
--cursor Open Cursor IDE with prompt on clipboard
--windsurf Open Windsurf IDE with prompt on clipboard
--clipboard Copy translation prompt to clipboard only
--stdout Print the translation prompt to stdout and exit, launching nothing (for an agent already in session)
--feature <names> Filter to specific feature(s) (comma-separated)
--context <file.json> Version 1 per-threat authoring context pack
--contract <contract> Template output contract: scanner (default) or probe
--from <path> Convert a scanner template file or directory to draft JS probes (repeatable; requires –contract probe) (default: [])
--headless Run an author CLI to completion without a TTY
--agent <agent> Headless author: claude, codex or gemini
--agent-timeout <ms> Headless author timeout in milliseconds (default 900000)
--out <dir> Template output directory (default .guardlink/cxg-templates/)
--threat <id> Select a threat or exposure id (repeatable) (default: [])
--format <format> Headless result format: json
--no-code-graph Do not supplement translate with optional graph facts
--code-graph-timeout <ms> Graph query timeout in milliseconds (default 300000; also GUARDLINK_CODEGRAPH_TIMEOUT_MS)
-h, --help display help for command

Merge multiple repo report JSONs into a unified workspace threat model

guardlink merge [options] <files...>

Arguments

Argument Description
files Report JSON file paths. Globs are expanded here, so quote them: ‘*/guardlink-report.json’

Options

Option Description
-o, --output <file> Output file for merged dashboard HTML (default: workspace-dashboard.html)
--json <file> Also write merged report JSON to this file
--diff-against <file> Compare against a previous merged JSON for weekly summary
-w, --workspace <name> Workspace name (auto-detected from reports if not set)
--summary-only Print only the text summary, skip dashboard generation
--strict Exit 1 when the estate has unmitigated exposures, confirmed exploits, or a repository that could not be read. Off by default
-h, --help display help for command

Estate-wide open risk, read back from the merged report guardlink merge --json wrote (exit 0 unless –strict)

guardlink estate [options] [file]

Arguments

Argument Description
file Merged report JSON (default: “workspace-merge.json”)

Options

Option Description
-f, --format <fmt> Output format: text (default) or json (default: “text”)
--severity <levels> Report only findings at these severities (comma-separated: critical,high,medium,low). Unrated findings always count
--repo <names> Report only findings from these repos (comma-separated)
--strict Exit 1 when anything is open or confirmed, or when the estate was not fully read
-h, --help display help for command

Collect the selected application semantic model and versioned completeness/revision receipt

guardlink application-view [options] [dir]

Arguments

Argument Description
dir Application root (default: “.”)

Options

Option Description
--manifest <file> Selected member manifest (JSON)
--member <id> Project the selected member only
--capabilities Print contract capabilities without reading a project
--json Print JSON (the default)
--strict Fail if mandatory selected inputs are incomplete
-h, --help display help for command

Link repos into a shared workspace for cross-repo threat modeling

guardlink link-project [options] [repos...]

Arguments

Argument Description
repos Repo directories to link (fresh setup: 2+ paths)

Options

Option Description
-w, --workspace <n> Workspace name (fresh link only) (default: “workspace”)
-r, --registry <url> GitHub/GitLab org base URL (e.g. github.com/unstructured)
--member <name=path> Explicit member name and directory (repeatable, fresh link) (default: [])
--name <name> Explicit member name for –add
--add <path> Add a new repo to an existing workspace (provide path to new repo)
--remove <name> Remove a repo from the workspace by name
--from <path> Existing workspace repo to read config from (used with –add or –remove)
-h, --help display help for command

Manage and inspect feature tags across the threat model

guardlink feature [options] [command]

Subcommands

Subcommand Summary
list List all features found in annotations
show Show detailed threat model for a specific feature

Options

Option Description
-h, --help display help for command

List all features found in annotations

guardlink feature list [options] [dir]

Arguments

Argument Description
dir Project directory to scan (default: “.”)

Options

Option Description
-p, --project <n> Project name (default: the name in .guardlink/config.json)
--json Output as JSON
-h, --help display help for command

Show detailed threat model for a specific feature

guardlink feature show [options] <name>

Arguments

Argument Description
name Feature name (case-insensitive)

Options

Option Description
-d, --dir <dir> Project directory (default: “.”)
-p, --project <n> Project name (default: the name in .guardlink/config.json)
--json Output as JSON
-h, --help display help for command