| name | external-software-analysis |
| description | External software analysis workflows for binaries/packages (decompilation, reverse engineering, static analysis, fuzzing, and evidence collection). Use when analyzing third-party software without source access. |
External Software Analysis
Overview
Guide external software analysis from decompilation to vulnerability discovery and evidence collection. Output is a structured Markdown report.
Workflow
- Read
references/external_sources.md for local canonical references.
- Apply the discovery process (decompile -> static scan -> flow tracing -> evidence).
- Produce a Markdown report following the reference report structure.
- When summarizing, map severity using
references/severity_criteria.md and references/severity_criteria_detail.md.
- If JSON findings are emitted, they must conform to
schemas/finding_schema.json and include:
provenance (one of binary-confirmed|source-confirmed|runtime-confirmed|not-confirmed)
impacted_flow (one or more flow IDs such as F1, F2)
Skill Interop: security-architecture-review
- Use this handoff when binary/package findings affect service trust boundaries, data flows, or authz decisions.
- Produce a compact handoff section (or separate markdown) with:
- External component name/version/artifact path
- Integration points (caller/callee, endpoint/function, file evidence)
- Data-in/data-out and sensitive fields touched
- Security control assumptions at the boundary (TLS, signature, token validation, pinning)
- Finding-to-flow mapping (
finding_id -> impacted flow/boundary)
- Evidence provenance tag per claim:
binary-confirmed, source-confirmed, runtime-confirmed, not-confirmed
- When source code is available for the integrating service, explicitly cross-check and upgrade provenance tags where possible.
- Reference template:
references/architecture_handoff.md.
Reporting
- Primary output: Markdown report (example reference provided).
- Use severity mapping from
references/severity_criteria.md and detailed criteria in references/severity_criteria_detail.md.
- Produce a common summary JSON using
schemas/reporting_summary_schema.json in this skill directory.
- When task-level JSON is required, use
schemas/task_output_schema.json.
- Use
scripts/generate_reporting_summary.py in this skill directory to build JSON output.
- For cross-skill usage, also emit a handoff markdown file (default:
./external-analysis-architecture-handoff.md) using references/architecture_handoff.md.
Resources
references/
references/external_sources.md
references/external_report_template.md
references/discovery_process.md
references/severity_criteria.md
references/severity_criteria_detail.md
references/reporting_summary.md
references/architecture_handoff.md
schemas/
schemas/reporting_summary_schema.json
schemas/task_output_schema.json
schemas/finding_schema.json
scripts/
scripts/generate_reporting_summary.py