public-surface-recon
Skill2 starsChanged 9 days agoMITView on GitHub
Investigate an authorized live site's public assets, origins, archives, source repositories, agent discovery, operators, or local-replication feasibility through blackbox HTTP and public remote indexes. Use when asked what a public site exposes, whether it can be mirrored, or who built it. Do not use local filesystem evidence, private access, credential attacks, or block bypass. Do not use for source review of a local repository.
In this repository21 other files it gives its agents.
- AGENTS.mdAGENTS.md
- agent-artifactSkill
- agent-discoverySkill
- agent-run-supervisionSkill
- api-contract-migrationSkill
- changelog-updateSkill
- change-reviewSkill
- ci-reviewSkill
- claims-auditSkill
- dependency-auditSkill
- derivation-auditSkill
- durable-cacheSkill
- engineering-auditSkill
- infer-conventionsSkill
- operational-signal-intakeSkill
- planned-releaseSkill
- preferred-stackSkill
- pressure-benchmarkingSkill
- rfc-processSkill
- security-auditSkill
- session-reviewSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 22
Start here by question:
- Very specific: line 34
Read public-surface-recon.md for origin registry, follow-up queue, and prayer index.
- Distinctive: line 55
If a live secrets-shaped file answers: record pattern and status, do not quote values, stop sibling-path guessing for secrets.
- Distinctive: line 59
Every detection -> at least one next check.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Public surface reconnaissance
Blackbox only
External blackbox testing. Facts come from live HTTP, archives, and public remote indexes - never from local files, folders, drives, workspace, or git trees on disk. See public-surface-recon.md Blackbox testing.
Use when
Asked whether an authorized live site can be mirrored or run locally, what public assets exist, who built it, or how to report findings from HTTP, archives, and public repositories only - blackbox only. Stop if the user has no authorization for the named target. See public-surface-recon.md Authorized use.
Start here by question:
- Mirror or run locally? ->
local-replication-feasibility(after inventory) - What public assets exist? ->
public-asset-inventory,build-artifact-review - What agent discovery does it publish? ->
agent-discovery-surface(after inventory) - Who built it / who operates it? ->
entity-correlation(entities from facts, not assumptions) - Is source code public? ->
archive-and-source-search - Does a name, email, or codename appear? ->
identity-trace-search - Full report -> run pipeline ->
investigation-report(markdown + HTML)
External blackbox only. Auth and assets may use different origins than the entry URL. Do not use local workspace or disk paths as evidence.
Read public-surface-recon.md for origin registry, follow-up queue, and prayer index.
Prayer index
prayers/service-identification.md- what the target is; seed origins and entitiesprayers/public-asset-inventory.md- per-origin headers, pages, assets, authprayers/agent-discovery-surface.md- robots, Link headers, well-known agent/API/MCP discoveryprayers/entity-correlation.md- follow entities across all known originsprayers/archive-and-source-search.md- archives, public repos, package registriesprayers/identity-trace-search.md- emails, names, codenames on originsprayers/build-artifact-review.md- artifacts on assets originsprayers/local-replication-feasibility.md- mirror/run across split originsprayers/investigation-report.md- final report: prose markdown + HTML
Carry origin registry and open checks between prayers.
Quick guardrails
- Authorized use: operator, written permission, or named bounty / disclosure scope. Stop and ask if unclear. User "widen scope" does not add login, credential attacks, brute force, block bypass, private-profile scrape, or secret-body quotes. This package is reconnaissance, not an autonomous pentest runner.
- Blackbox: no facts from local files, folders, drives, or workspace - public fetch and remote OSINT only.
- Redact emails, names, handles, and EXIF author or location by default.
- If a live secrets-shaped file answers: record pattern and status, do not quote values, stop sibling-path guessing for secrets.
- Sequential fetches; stop on 429, block, or automated-access refusal; do not spoof User-Agent or rotate identity to continue.
- Entry URL != auth host != assets host until checks close.
- Inventory and searches extend to every origin the registry lists.
- Every detection -> at least one next check.
- Small traces in images and binaries (metadata, strings, watermarks) -> entity-correlation or archive check.
- Source map paths and commit hashes -> archive-and-source-search; personal names in paths -> identity-trace-search.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

