using-ecosystem

Name: using-ecosystem
Author: krzysztofsurdy

16installs1forks

How to install using-ecosystem?

krzysztofsurdy/code-virtuoso/using-ecosystem

$npx -y skills add krzysztofsurdy/code-virtuoso --skill using-ecosystem

Installs into the current project.

Summary

Ecosystem discovery advisor. Use when the user asks 'what skill should I use', 'what agent should I delegate to', 'which team fits this task', or when onboarding to available skills, agents, and teams. Scans ALL installed skills at runtime - not limited to any single plugin or ve

Files · 1

View on GitHub

SKILL.md

1	# Using Ecosystem
2
3	Discover the right skill, agent, or team for the situation at hand. This skill teaches a discovery PROCESS - it does not maintain a hardcoded inventory. Every recommendation comes from scanning what is actually installed.
4
5	## Core Principles
6
7	\| Principle \| Meaning \|
8	\|---\|---\|
9	\| Scan, don't guess \| Always scan installed skills/agents/teams before recommending. Never assume something is available. \|
10	\| Match situation, not topic \| Pick the skill whose description triggers match what the user is doing, not the skill whose title loosely matches. \|
11	\| Agents act, skills inform \| Skills are reference material. Agents are actors. Delegate work to an agent; consult a skill. \|
12	\| Teams for multi-role work \| If the work needs 3+ agents coordinating, check for a pre-composed team before assembling ad-hoc. \|
13	\| Narrower wins \| If two skills overlap, pick the one with the tighter trigger match. \|
14	\| Chain when work crosses domains \| Multi-step tasks often need two or three agents in sequence. \|
15
16	---
17
18	## Step 1: Scan What Is Installed
19
20	Run the discovery commands from [discovery-commands](references/discovery-commands.md) to build a live index of installed skills, agents, and teams. Read each frontmatter `name` and `description` field.
21
22	Do NOT skip this step. Do NOT recommend from memory. The installed set varies per user and per project.
23
24	---
25
26	## Step 2: Classify the User's Situation
27
28	Determine what the user needs:
29
30	\| Signal \| Category \|
31	\|---\|---\|
32	\| Needs reference material, patterns, or principles \| Knowledge skill \|
33	\| Needs to generate a specific output (ticket, PR message, report, rules file) \| Tool skill \|
34	\| Needs to follow a step-by-step operational procedure \| Playbook skill \|
35	\| Needs framework-specific component reference \| Framework skill \|
36	\| Needs a single focused task done (investigate, review, implement) \| Specialist agent \|
37	\| Needs ongoing domain ownership (requirements, architecture, QA) \| Role agent \|
38	\| Needs coordinated multi-agent delivery (feature, release, review cycle) \| Team \|
39
40	---
41
42	## Step 3: Match and Recommend
43
44	### Decision Tree
45
46	```
47	User request
48	\|
49	v
50	Does it need multiple agents coordinating?
51	+-- Yes --> Scan bundled and project teams for a match
52	\| +-- Found --> Recommend the team
53	\| +-- None --> Assemble ad-hoc from agents (see chaining below)
54	\|
55	+-- No --> Is it a well-scoped actionable task?
56	+-- Yes --> Scan agents/ for a matching specialist or role
57	+-- No --> Scan skills/ for a matching skill
58	Use the category signals from Step 2
59	```
60
61	### Matching Rules
62
63	1. Read the `description` field of each candidate. Match on trigger phrases, not the name.
64	2. Prefer the narrower skill. `database-design` beats `clean-architecture` for a schema question.
65	3. Check `user-invocable` - if the user wants to run something interactively, only recommend user-invocable skills.
66	4. Skills stack. You can recommend multiple knowledge skills for one task (e.g., `security` + `solid` for a security review).
67	5. Never recommend what is not installed. If a skill would fit but is not present, suggest installing it with `npx skills add`.
68
69	---
70
71	## Step 4: Chain When Needed
72
73	When a single skill or agent is not enough, chain them. Common patterns:
74
75	\| Pattern \| Shape \|
76	\|---\|---\|
77	\| Investigation flow \| Read-only agent --> Design agent --> Implementation agent --> Review agent \|
78	\| Feature flow \| Requirements role --> Architecture role --> Dev roles (parallel) --> QA role \|
79	\| Review flow \| Smell scanner --> Reviewer --> (optional) Implementer for fixes \|
80	\| Coverage flow \| Test gap analyzer --> Implementer --> Reviewer \|
81
82	These are guidelines, not rituals. Skip steps that do not apply. Add steps that do. If a pre-composed team matches the chain, use the team instead.
83
84	---
85
86	## Step 5: Compass - What Next?
87
88	When the user has just finished work with an agent, skill, or team and asks "what next?" or "where do I go from here?":
89
90	1. Ask what they just completed (which agent/skill/team, what output they have)
91	2. Consult [compass](references/compass.md) for the recommended next step
92	3. Verify the recommended agent/skill/team is installed (scan first - do not recommend from memory)
93	4. Present the recommendation with one-line reasoning
94	5. If multiple paths exist, present 2-3 options and let the user choose
95	6. If nothing chains naturally, say so - not every tas

People who liked this also viewed

6 picks

Type

using-ecosystem

16installs1forks

View on GitHub

How to install using-ecosystem?

krzysztofsurdy/code-virtuoso/using-ecosystem

$npx -y skills add krzysztofsurdy/code-virtuoso --skill using-ecosystem

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Using Ecosystem
2
3	Discover the right skill, agent, or team for the situation at hand. This skill teaches a discovery PROCESS - it does not maintain a hardcoded inventory. Every recommendation comes from scanning what is actually installed.
4
5	## Core Principles
6
7	\| Principle \| Meaning \|
8	\|---\|---\|
9	\| Scan, don't guess \| Always scan installed skills/agents/teams before recommending. Never assume something is available. \|
10	\| Match situation, not topic \| Pick the skill whose description triggers match what the user is doing, not the skill whose title loosely matches. \|
11	\| Agents act, skills inform \| Skills are reference material. Agents are actors. Delegate work to an agent; consult a skill. \|
12	\| Teams for multi-role work \| If the work needs 3+ agents coordinating, check for a pre-composed team before assembling ad-hoc. \|
13	\| Narrower wins \| If two skills overlap, pick the one with the tighter trigger match. \|
14	\| Chain when work crosses domains \| Multi-step tasks often need two or three agents in sequence. \|
15
16	---
17
18	## Step 1: Scan What Is Installed
19
20	Run the discovery commands from [discovery-commands](references/discovery-commands.md) to build a live index of installed skills, agents, and teams. Read each frontmatter `name` and `description` field.
21
22	Do NOT skip this step. Do NOT recommend from memory. The installed set varies per user and per project.
23
24	---
25
26	## Step 2: Classify the User's Situation
27
28	Determine what the user needs:
29
30	\| Signal \| Category \|
31	\|---\|---\|
32	\| Needs reference material, patterns, or principles \| Knowledge skill \|
33	\| Needs to generate a specific output (ticket, PR message, report, rules file) \| Tool skill \|
34	\| Needs to follow a step-by-step operational procedure \| Playbook skill \|
35	\| Needs framework-specific component reference \| Framework skill \|
36	\| Needs a single focused task done (investigate, review, implement) \| Specialist agent \|
37	\| Needs ongoing domain ownership (requirements, architecture, QA) \| Role agent \|
38	\| Needs coordinated multi-agent delivery (feature, release, review cycle) \| Team \|
39
40	---
41
42	## Step 3: Match and Recommend
43
44	### Decision Tree
45
46	```
47	User request
48	\|
49	v
50	Does it need multiple agents coordinating?
51	+-- Yes --> Scan bundled and project teams for a match
52	\| +-- Found --> Recommend the team
53	\| +-- None --> Assemble ad-hoc from agents (see chaining below)
54	\|
55	+-- No --> Is it a well-scoped actionable task?
56	+-- Yes --> Scan agents/ for a matching specialist or role
57	+-- No --> Scan skills/ for a matching skill
58	Use the category signals from Step 2
59	```
60
61	### Matching Rules
62
63	1. Read the `description` field of each candidate. Match on trigger phrases, not the name.
64	2. Prefer the narrower skill. `database-design` beats `clean-architecture` for a schema question.
65	3. Check `user-invocable` - if the user wants to run something interactively, only recommend user-invocable skills.
66	4. Skills stack. You can recommend multiple knowledge skills for one task (e.g., `security` + `solid` for a security review).
67	5. Never recommend what is not installed. If a skill would fit but is not present, suggest installing it with `npx skills add`.
68
69	---
70
71	## Step 4: Chain When Needed
72
73	When a single skill or agent is not enough, chain them. Common patterns:
74
75	\| Pattern \| Shape \|
76	\|---\|---\|
77	\| Investigation flow \| Read-only agent --> Design agent --> Implementation agent --> Review agent \|
78	\| Feature flow \| Requirements role --> Architecture role --> Dev roles (parallel) --> QA role \|
79	\| Review flow \| Smell scanner --> Reviewer --> (optional) Implementer for fixes \|
80	\| Coverage flow \| Test gap analyzer --> Implementer --> Reviewer \|
81
82	These are guidelines, not rituals. Skip steps that do not apply. Add steps that do. If a pre-composed team matches the chain, use the team instead.
83
84	---
85
86	## Step 5: Compass - What Next?
87
88	When the user has just finished work with an agent, skill, or team and asks "what next?" or "where do I go from here?":
89
90	1. Ask what they just completed (which agent/skill/team, what output they have)
91	2. Consult [compass](references/compass.md) for the recommended next step
92	3. Verify the recommended agent/skill/team is installed (scan first - do not recommend from memory)
93	4. Present the recommendation with one-line reasoning
94	5. If multiple paths exist, present 2-3 options and let the user choose
95	6. If nothing chains naturally, say so - not every tas

People who liked this also viewed

6 picks

Type