MCP Tools Reference
All 16 tools available in @squirex.dev/mcp-server. Parameter names below
match the server's tools/list input schemas exactly.
Results and errors. Scan tools return the SARIF 2.1.0 log as the first
text item and a one-line summary as the second. Findings are results, not
errors (the engine exits 2 when it finds violations). isError: true is set
only when a tool could not run โ CLI or engine missing, invalid arguments, no
output โ and the text explains why.
Annotations. Every tool declares a title and MCP behavior hints, so
clients can decide what needs confirmation:
| Hint set | Tools |
|---|---|
Read-only, local (readOnlyHint: true, openWorldHint: false) | scan_agentforce, scan_agentforce_file, scan_agentforce_rule, list_scan_rules, get_rule_details, explain_violation, suggest_fix, predict_conflicts |
Local side effects (readOnlyHint: false, destructiveHint: false) | run_tests, get_coverage, analyze_schema, generate_sarif_report, generate_dx_tests, validate_dx_tests |
Read-only, external (openWorldHint: true) | get_testing_center_results |
Destructive, external (destructiveHint: true, openWorldHint: true) | push_to_testing_center |
Core Scanningโ
scan_agentforceโ
Full project scan (squirex scan). With baseBranch, a PR-scoped scan of
changed files only (squirex scan-pr --base).
Parameters:
directory(optional) โ directory to scan, relative toSQUIREX_PROJECT_DIRor absolutebaseBranch(optional) โ diff against this branch; omit for a full scan
scan_agentforce_fileโ
Findings for a single file. Many rules are cross-file, so the server scans the whole project (or the file's directory if it is outside the project) and returns only the SARIF results located in this file.
Parameters:
filePathโ path to the file, relative to the project root or absolute
scan_agentforce_ruleโ
Runs a single rule against the full project (--rules).
Parameters:
ruleIdโ e.g.AGENTFORCE-1.1
Rule Intelligenceโ
list_scan_rulesโ
Returns rules with id, category, severity, and short description.
Parameters:
category(optional)severity(optional)
get_rule_detailsโ
Returns full rule documentation including remediation guidance.
Parameters:
ruleId
explain_violationโ
Rule-level root-cause explanation for a SARIF result.
Parameters:
ruleIdfilePath(optional)message(optional) โ the SARIF result messagelineNumber(optional)
suggest_fixโ
Rule-level fix template for a violation (does not read the file).
Parameters:
ruleIdfilePathviolationContext(optional) โ the violating XML/Apex snippet
Apex Testing & Schemaโ
run_testsโ
Executes Apex tests using the Go runtime engine. Failing tests are results, not tool errors.
Parameters:
file(optional) โ test file, e.g.MyTest.clsfilter(optional) โ regex for test method namesparallel(optional, boolean)failFast(optional, boolean)
get_coverageโ
Returns line-level coverage data (runs the tests).
Parameters:
className(optional)
analyze_schemaโ
Returns the inferred SObject schema from the codebase.
Parameters:
directory(optional) โ default./force-appincludeHeatmap(optional, boolean)live(optional, boolean, defaulttrue) โ aggregate live via the Go engine; writesschema_graph.jsonin the project
predict_conflictsโ
Parameters:
branchesโ comma-separated branch names (e.g.main,feature/my-branch)simulate(optional, boolean)
generate_sarif_reportโ
Runs a scan and returns the SARIF log, or writes it to outputPath.
Parameters:
outputPath(optional) โ file to write, relative to the project directorybaseBranch(optional) โ PR-scoped report
Testing Center Bridgeโ
generate_dx_testsโ
Converts scan violations to Agentforce DX test YAML.
Parameters:
sarifPath(optional) โ existing SARIF file; a fresh scan runs if omittedoutputPath(optional)suiteName(optional)rules(optional) โ comma-separated rule IDs
validate_dx_testsโ
Validates a generated DX test spec.
Parameters:
testFilePath
push_to_testing_centerโ
Pushes the spec to Salesforce via sf agent test run.
Parameters:
testFilePathtargetOrgโ authenticated org alias
get_testing_center_resultsโ
Returns instructions for checking a Testing Center run.
Parameters:
testRunIdtargetOrg