quickstart

Name: quickstart
Author: anthropics

6.7kstars543forks

How to install quickstart?

anthropics/defending-code-reference-harness/quickstart

$npx -y skills add anthropics/defending-code-reference-harness --skill quickstart

Installs into the current project.

Summary

The front door for this repo. With no argument: a 30-second intro, then an offer to walk you through your first run on the canary target. With a question: answers it from this repo's own docs and source, cites where it looked, and hands you the next command. Use for "how do I…",

Files · 1

View on GitHub

SKILL.md

1	# /quickstart
2
3	Two modes, picked by whether `$ARGUMENTS` is empty.
4
5	- Empty → Intro mode. Short orientation, then offer the guided first run.
6	- Non-empty → Help mode. Treat `$ARGUMENTS` as the operator's question.
7
8	---
9
10	## Intro mode
11
12	Keep it short and a little warm; this is the first thing a new operator sees.
13
14	Say roughly:
15
16	> Welcome! This repo takes you from finding your first vulnerability to
17	> patching at scale, using a set of Claude Code skills and an autonomous
18	> pipeline. Two ways in: interactive skills (no setup, safe, start here)
19	> and the autonomous pipeline (Docker, scales to hundreds of parallel
20	> agents).
21	>
22	> The ramp-up:
23	>
24	> \| Day 1 \| Threat-model + first static scan + triage \|
25	> \| Day 2 \| Run the reference pipeline (C/C++) \|
26	> \| Day 3-4 \| Customize it for your stack \|
27	> \| Week 2 \| Autonomous scanning, triage, and patching \|
28	>
29	> Day-1 goal: threat-model, scan, and triage the bundled canary target.
30	> Most teams get there before lunch.
31
32	Remind them to `export CLAUDE_CODE_SUBAGENT_MODEL=<model-id>` so subagents
33	use the same model as the session.
34
35	Then AskUserQuestion with three options:
36
37	1. Walk me through Day 1 on the canary (~10 min) → run "Guided first
38	run" below.
39	2. I have a question → ask what it is, then switch to Help mode.
40	3. I'll read the README → point at `README.md` Step 1 and stop.
41
42	### Guided first run
43
44	Runs the three Step-1 skills on `targets/canary`, pausing after each to show
45	what landed on disk. These only read/write files in the repo; no sandbox
46	needed.
47
48	1. `/threat-model bootstrap targets/canary` via Task. When done, open
49	`THREAT_MODEL.md`, show the focus areas, explain in 2-3 sentences how
50	this steers the scan.
51	2. `/vuln-scan targets/canary` via Task. When done, open
52	`targets/canary/VULN-FINDINGS.md`, summarize the count and top 2-3
53	findings, point at `VULN-FINDINGS.json`.
54	3. `/triage targets/canary/VULN-FINDINGS.json` via Task. When done, open
55	`TRIAGE.md`, explain what changed vs. raw findings (verified, deduped,
56	re-ranked).
57
58	Pause for the operator between each (AskUserQuestion); don't barrel through.
59	Close with a one-line recap of the three artifacts on disk, then point at
60	README Step 2 for the execution-verified pipeline. **Never run `vuln-pipeline`
61	or anything that executes target code here**; that's Step 2 and needs
62	Docker + a sandbox.
63
64	---
65
66	## Help mode
67
68	Answer the operator's question using this repo as ground truth: README,
69	`docs/.md`, `harness/.py`, `dnr_harness/.py`, `targets//config.yaml`,
70	`.claude/skills/*`.
71	Don't answer from general knowledge when the repo has a specific answer.
72
73	### Routing map
74
75	\| If the question is about… \| Read first \| Then offer \|
76	\|---------------------------------\|-----------------------------------------\|------------\|
77	\| running the pipeline \| `docs/pipeline.md`, README Step 2 \| the `recon` / `run` command \|
78	\| too many findings, triage \| `docs/triage.md` \| `/triage <path>` \|
79	\| porting, Java/Go/Rust/etc. \| `docs/customizing.md`, README Step 3 \| `/customize` \|
80	\| safety, sandbox, Docker \| `docs/security.md` \| cite; no action \|
81	\| rate limits, 429, token budget \| `docs/pipeline.md`: Rate limits, `docs/troubleshooting.md#rate-limits` \| cite the numbers \|
82	\| duplicates, dedup \| `docs/troubleshooting.md#duplicate-findings` \| `known_bugs:` hint \|
83	\| CLI flags, "what does --X do" \| `harness/cli.py` (grep the argparse) \| exact flag + example \|
84	\| which model, subagent pinning \| `docs/troubleshooting.md`: Subagents \| the `export` line \|
85	\| best practices, prompting \| `docs/best-practices.md`, `docs/prompting.md` \| cite the principle \|
86	\| threat model, attack surface, scope \| `docs/threat-model.md` \| `/threat-model bootstrap <target-dir>` \|
87	\| scan, audit, find vulns \| `.claude/skills/vuln-scan/SKILL.md` \| `/vuln-scan <target-dir>` \|
88	\| "how do I start" \| README Step 1 \| offer Guided first run \|
89	\| patching, fix, diff, re-attack \| `docs/patching.md`, README Step 4 \| `/patch <input>` \|
90	\| threat hunting, incident response, logs \| `docs/detection-response.md` \| `/dnr-hunt` or `/dnr-respond` \|
91	\| autonomous D&R, dnrcanary \| `docs/detection-response.md`, `targets/dnrcanary/README.md` \| the `dnr-pipeline run` command \|
92	\|

People who liked this also viewed

6 picks

Type

quickstart

6.7kstars543forks

View on GitHub

How to install quickstart?

anthropics/defending-code-reference-harness/quickstart

$npx -y skills add anthropics/defending-code-reference-harness --skill quickstart

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# /quickstart
2
3	Two modes, picked by whether `$ARGUMENTS` is empty.
4
5	- Empty → Intro mode. Short orientation, then offer the guided first run.
6	- Non-empty → Help mode. Treat `$ARGUMENTS` as the operator's question.
7
8	---
9
10	## Intro mode
11
12	Keep it short and a little warm; this is the first thing a new operator sees.
13
14	Say roughly:
15
16	> Welcome! This repo takes you from finding your first vulnerability to
17	> patching at scale, using a set of Claude Code skills and an autonomous
18	> pipeline. Two ways in: interactive skills (no setup, safe, start here)
19	> and the autonomous pipeline (Docker, scales to hundreds of parallel
20	> agents).
21	>
22	> The ramp-up:
23	>
24	> \| Day 1 \| Threat-model + first static scan + triage \|
25	> \| Day 2 \| Run the reference pipeline (C/C++) \|
26	> \| Day 3-4 \| Customize it for your stack \|
27	> \| Week 2 \| Autonomous scanning, triage, and patching \|
28	>
29	> Day-1 goal: threat-model, scan, and triage the bundled canary target.
30	> Most teams get there before lunch.
31
32	Remind them to `export CLAUDE_CODE_SUBAGENT_MODEL=<model-id>` so subagents
33	use the same model as the session.
34
35	Then AskUserQuestion with three options:
36
37	1. Walk me through Day 1 on the canary (~10 min) → run "Guided first
38	run" below.
39	2. I have a question → ask what it is, then switch to Help mode.
40	3. I'll read the README → point at `README.md` Step 1 and stop.
41
42	### Guided first run
43
44	Runs the three Step-1 skills on `targets/canary`, pausing after each to show
45	what landed on disk. These only read/write files in the repo; no sandbox
46	needed.
47
48	1. `/threat-model bootstrap targets/canary` via Task. When done, open
49	`THREAT_MODEL.md`, show the focus areas, explain in 2-3 sentences how
50	this steers the scan.
51	2. `/vuln-scan targets/canary` via Task. When done, open
52	`targets/canary/VULN-FINDINGS.md`, summarize the count and top 2-3
53	findings, point at `VULN-FINDINGS.json`.
54	3. `/triage targets/canary/VULN-FINDINGS.json` via Task. When done, open
55	`TRIAGE.md`, explain what changed vs. raw findings (verified, deduped,
56	re-ranked).
57
58	Pause for the operator between each (AskUserQuestion); don't barrel through.
59	Close with a one-line recap of the three artifacts on disk, then point at
60	README Step 2 for the execution-verified pipeline. **Never run `vuln-pipeline`
61	or anything that executes target code here**; that's Step 2 and needs
62	Docker + a sandbox.
63
64	---
65
66	## Help mode
67
68	Answer the operator's question using this repo as ground truth: README,
69	`docs/.md`, `harness/.py`, `dnr_harness/.py`, `targets//config.yaml`,
70	`.claude/skills/*`.
71	Don't answer from general knowledge when the repo has a specific answer.
72
73	### Routing map
74
75	\| If the question is about… \| Read first \| Then offer \|
76	\|---------------------------------\|-----------------------------------------\|------------\|
77	\| running the pipeline \| `docs/pipeline.md`, README Step 2 \| the `recon` / `run` command \|
78	\| too many findings, triage \| `docs/triage.md` \| `/triage <path>` \|
79	\| porting, Java/Go/Rust/etc. \| `docs/customizing.md`, README Step 3 \| `/customize` \|
80	\| safety, sandbox, Docker \| `docs/security.md` \| cite; no action \|
81	\| rate limits, 429, token budget \| `docs/pipeline.md`: Rate limits, `docs/troubleshooting.md#rate-limits` \| cite the numbers \|
82	\| duplicates, dedup \| `docs/troubleshooting.md#duplicate-findings` \| `known_bugs:` hint \|
83	\| CLI flags, "what does --X do" \| `harness/cli.py` (grep the argparse) \| exact flag + example \|
84	\| which model, subagent pinning \| `docs/troubleshooting.md`: Subagents \| the `export` line \|
85	\| best practices, prompting \| `docs/best-practices.md`, `docs/prompting.md` \| cite the principle \|
86	\| threat model, attack surface, scope \| `docs/threat-model.md` \| `/threat-model bootstrap <target-dir>` \|
87	\| scan, audit, find vulns \| `.claude/skills/vuln-scan/SKILL.md` \| `/vuln-scan <target-dir>` \|
88	\| "how do I start" \| README Step 1 \| offer Guided first run \|
89	\| patching, fix, diff, re-attack \| `docs/patching.md`, README Step 4 \| `/patch <input>` \|
90	\| threat hunting, incident response, logs \| `docs/detection-response.md` \| `/dnr-hunt` or `/dnr-respond` \|
91	\| autonomous D&R, dnrcanary \| `docs/detection-response.md`, `targets/dnrcanary/README.md` \| the `dnr-pipeline run` command \|
92	\|

People who liked this also viewed

6 picks

Type