alphacouncil-agent 0.5.0 → 0.5.2

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (11) hide show
  1. package/.claude/commands/alpha.md +79 -0
  2. package/.claude-plugin/marketplace.json +2 -2
  3. package/.claude-plugin/plugin.json +3 -2
  4. package/.codex-plugin/plugin.json +1 -1
  5. package/.grok/commands/alpha.md +79 -0
  6. package/.opencode/command/alpha.md +79 -0
  7. package/AGENTS.md +26 -0
  8. package/CHANGELOG.md +27 -0
  9. package/README.md +22 -0
  10. package/commands/alpha.md +79 -0
  11. package/package.json +6 -2
package/.claude/commands/alpha.md ADDED
@@ -0,0 +1,79 @@
1
+ ---
2
+ description: Equity research council — full run, quick read, mechanical screen, or market narrative
3
+ argument-hint: [ticker] [quick|screen|market|options|news] · or just a question
4
+ ---
5
+
6
+ # /alpha
7
+
8
+ Request: **$ARGUMENTS**
9
+
10
+ ## Route on the arguments — do not ask what they meant
11
+
12
+ | `$ARGUMENTS` looks like | Do this |
13
+ |---|---|
14
+ | *(empty)* | Print the mode table below and stop. Do not start a run. |
15
+ | a ticker alone, or a question | **Full council.** Go to "Full council" below. |
16
+ | ticker + `quick` | The `quick` preset: 4 analysts and the debate. No bench, no verification. |
17
+ | ticker + `screen` | `screen_ticker` only. No language-model judgment, no subagents. |
18
+ | ticker + `options` | `get_options_chain` only. |
19
+ | ticker + `news` | `get_news` on the symbol, and on its CIK when it is a US filer. |
20
+ | `market` (with or without a theme) | `get_market_narrative`. Add the theme as `extra_queries`. |
21
+
22
+ When `$ARGUMENTS` is empty, print exactly this and stop:
23
+
24
+ ```
25
+ /alpha MU full council — asks which preset first
26
+ /alpha MU quick 4 analysts + debate, no bench, no verification
27
+ /alpha MU screen mechanical filings screen only (no model spend)
28
+ /alpha MU options IV term structure, skew, positioning (no model spend)
29
+ /alpha MU news dated filings and headlines (no model spend)
30
+ /alpha market AI what the market is talking about (no model spend)
31
+ ```
32
+
33
+ Say plainly that the four marked modes call keyless data tools and spawn no subagents, so
34
+ they cost nothing beyond this turn. The council modes spawn one subagent per seat, and that
35
+ is where the spend is.
36
+
37
+ ## Full council
38
+
39
+ 1. **Ask which preset first.** Call `list_council_options` and present the three with their
40
+ seat counts. A council is 7 to 44 seats and that range is the user's time and money.
41
+ **Skip the question when the request already answered it** — a named roster, "everything",
42
+ "be quick". A confirmation nobody needed is an interruption.
43
+ 2. Follow `skills/alphacouncil-agent/SKILL.md` from Stage 0. Do not improvise a shorter
44
+ workflow: a report that looks finished and skipped the bench is worse than an obviously
45
+ partial one.
46
+ 3. **Call the tools rather than searching for numbers they can supply.** `screen_ticker`,
47
+ `get_quote`, `get_options_chain`, `get_macro_snapshot`, `get_news`. Search is for
48
+ explanation and for what is not yet filed.
49
+ 4. Every selected master must report before the run is complete. `out_of_scope` is a
50
+ conclusion, not an abstention.
51
+ 5. Give price bands with the condition attached to each. "The cycle position is undetermined"
52
+ changes what the bands depend on; it does not excuse leaving them out.
53
+
54
+ ## Quick mode owes the reader one sentence
55
+
56
+ State in the output that no master lenses ran and nothing was cross-verified, so the
57
+ confidence is lower than a standard run. Without it, a four-seat run reads like a full
58
+ council result.
59
+
60
+ ## Screen mode
61
+
62
+ A rule whose inputs are missing is `skipped`, **never a pass** — reporting "6/7 passed"
63
+ without naming the seventh misrepresents the screen. Surviving is not a recommendation; it
64
+ means the name is worth research time. For a non-US ticker, say which market it is and use
65
+ `market_coverage` rather than returning nothing.
66
+
67
+ ## Market mode
68
+
69
+ Read it by the gap, not the ranking. A theme leading coverage while its series has not moved
70
+ means the story is ahead of the data, or the market stopped listening — say which. A series
71
+ that moved sharply while its theme is barely covered is the more valuable finding. Headline
72
+ counts measure attention, never truth.
73
+
74
+ ## Never
75
+
76
+ - Never fill a number from memory when a tool could have supplied it.
77
+ - Never let social or narrative evidence enter the conclusion on its own.
78
+
79
+ Educational software. The report says so.
package/.claude-plugin/marketplace.json CHANGED
@@ -6,14 +6,14 @@
6
6
  },
7
7
  "metadata": {
8
8
  "description": "AlphaCouncil plugins for public-equity research.",
9
- "version": "0.5.0"
9
+ "version": "0.5.2"
10
10
  },
11
11
  "plugins": [
12
12
  {
13
13
  "name": "alphacouncil-agent",
14
14
  "source": "./",
15
15
  "description": "Multi-agent equity research: evidence packets, bull/bear debate, portfolio-manager decision.",
16
- "version": "0.5.0",
16
+ "version": "0.5.2",
17
17
  "author": {
18
18
  "name": "Zhao73"
19
19
  },
package/.claude-plugin/plugin.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "alphacouncil-agent",
3
- "version": "0.5.0",
3
+ "version": "0.5.2",
4
4
  "description": "Multi-agent public-equity research workflow for Claude Code: spawns analyst workers, gathers sourced JSON evidence packets, runs bull/bear debate, and produces a portfolio-manager Buy/Overweight/Hold/Underweight/Sell decision with a full report.",
5
5
  "author": {
6
6
  "name": "Zhao73"
@@ -24,5 +24,6 @@
24
24
  "${CLAUDE_PLUGIN_ROOT}/mcp/server.mjs"
25
25
  ]
26
26
  }
27
- }
27
+ },
28
+ "commands": "./commands/"
28
29
  }
package/.codex-plugin/plugin.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "alphacouncil-agent",
3
- "version": "0.5.0",
3
+ "version": "0.5.2",
4
4
  "description": "Host-visible or background public-equity research workflow with shared evidence packets.",
5
5
  "author": {
6
6
  "name": "Zhao73"
package/.grok/commands/alpha.md ADDED
@@ -0,0 +1,79 @@
1
+ ---
2
+ description: Equity research council — full run, quick read, mechanical screen, or market narrative
3
+ argument-hint: [ticker] [quick|screen|market|options|news] · or just a question
4
+ ---
5
+
6
+ # /alpha
7
+
8
+ Request: **$ARGUMENTS**
9
+
10
+ ## Route on the arguments — do not ask what they meant
11
+
12
+ | `$ARGUMENTS` looks like | Do this |
13
+ |---|---|
14
+ | *(empty)* | Print the mode table below and stop. Do not start a run. |
15
+ | a ticker alone, or a question | **Full council.** Go to "Full council" below. |
16
+ | ticker + `quick` | The `quick` preset: 4 analysts and the debate. No bench, no verification. |
17
+ | ticker + `screen` | `screen_ticker` only. No language-model judgment, no subagents. |
18
+ | ticker + `options` | `get_options_chain` only. |
19
+ | ticker + `news` | `get_news` on the symbol, and on its CIK when it is a US filer. |
20
+ | `market` (with or without a theme) | `get_market_narrative`. Add the theme as `extra_queries`. |
21
+
22
+ When `$ARGUMENTS` is empty, print exactly this and stop:
23
+
24
+ ```
25
+ /alpha MU full council — asks which preset first
26
+ /alpha MU quick 4 analysts + debate, no bench, no verification
27
+ /alpha MU screen mechanical filings screen only (no model spend)
28
+ /alpha MU options IV term structure, skew, positioning (no model spend)
29
+ /alpha MU news dated filings and headlines (no model spend)
30
+ /alpha market AI what the market is talking about (no model spend)
31
+ ```
32
+
33
+ Say plainly that the four marked modes call keyless data tools and spawn no subagents, so
34
+ they cost nothing beyond this turn. The council modes spawn one subagent per seat, and that
35
+ is where the spend is.
36
+
37
+ ## Full council
38
+
39
+ 1. **Ask which preset first.** Call `list_council_options` and present the three with their
40
+ seat counts. A council is 7 to 44 seats and that range is the user's time and money.
41
+ **Skip the question when the request already answered it** — a named roster, "everything",
42
+ "be quick". A confirmation nobody needed is an interruption.
43
+ 2. Follow `skills/alphacouncil-agent/SKILL.md` from Stage 0. Do not improvise a shorter
44
+ workflow: a report that looks finished and skipped the bench is worse than an obviously
45
+ partial one.
46
+ 3. **Call the tools rather than searching for numbers they can supply.** `screen_ticker`,
47
+ `get_quote`, `get_options_chain`, `get_macro_snapshot`, `get_news`. Search is for
48
+ explanation and for what is not yet filed.
49
+ 4. Every selected master must report before the run is complete. `out_of_scope` is a
50
+ conclusion, not an abstention.
51
+ 5. Give price bands with the condition attached to each. "The cycle position is undetermined"
52
+ changes what the bands depend on; it does not excuse leaving them out.
53
+
54
+ ## Quick mode owes the reader one sentence
55
+
56
+ State in the output that no master lenses ran and nothing was cross-verified, so the
57
+ confidence is lower than a standard run. Without it, a four-seat run reads like a full
58
+ council result.
59
+
60
+ ## Screen mode
61
+
62
+ A rule whose inputs are missing is `skipped`, **never a pass** — reporting "6/7 passed"
63
+ without naming the seventh misrepresents the screen. Surviving is not a recommendation; it
64
+ means the name is worth research time. For a non-US ticker, say which market it is and use
65
+ `market_coverage` rather than returning nothing.
66
+
67
+ ## Market mode
68
+
69
+ Read it by the gap, not the ranking. A theme leading coverage while its series has not moved
70
+ means the story is ahead of the data, or the market stopped listening — say which. A series
71
+ that moved sharply while its theme is barely covered is the more valuable finding. Headline
72
+ counts measure attention, never truth.
73
+
74
+ ## Never
75
+
76
+ - Never fill a number from memory when a tool could have supplied it.
77
+ - Never let social or narrative evidence enter the conclusion on its own.
78
+
79
+ Educational software. The report says so.
package/.opencode/command/alpha.md ADDED
@@ -0,0 +1,79 @@
1
+ ---
2
+ description: Equity research council — full run, quick read, mechanical screen, or market narrative
3
+ argument-hint: [ticker] [quick|screen|market|options|news] · or just a question
4
+ ---
5
+
6
+ # /alpha
7
+
8
+ Request: **$ARGUMENTS**
9
+
10
+ ## Route on the arguments — do not ask what they meant
11
+
12
+ | `$ARGUMENTS` looks like | Do this |
13
+ |---|---|
14
+ | *(empty)* | Print the mode table below and stop. Do not start a run. |
15
+ | a ticker alone, or a question | **Full council.** Go to "Full council" below. |
16
+ | ticker + `quick` | The `quick` preset: 4 analysts and the debate. No bench, no verification. |
17
+ | ticker + `screen` | `screen_ticker` only. No language-model judgment, no subagents. |
18
+ | ticker + `options` | `get_options_chain` only. |
19
+ | ticker + `news` | `get_news` on the symbol, and on its CIK when it is a US filer. |
20
+ | `market` (with or without a theme) | `get_market_narrative`. Add the theme as `extra_queries`. |
21
+
22
+ When `$ARGUMENTS` is empty, print exactly this and stop:
23
+
24
+ ```
25
+ /alpha MU full council — asks which preset first
26
+ /alpha MU quick 4 analysts + debate, no bench, no verification
27
+ /alpha MU screen mechanical filings screen only (no model spend)
28
+ /alpha MU options IV term structure, skew, positioning (no model spend)
29
+ /alpha MU news dated filings and headlines (no model spend)
30
+ /alpha market AI what the market is talking about (no model spend)
31
+ ```
32
+
33
+ Say plainly that the four marked modes call keyless data tools and spawn no subagents, so
34
+ they cost nothing beyond this turn. The council modes spawn one subagent per seat, and that
35
+ is where the spend is.
36
+
37
+ ## Full council
38
+
39
+ 1. **Ask which preset first.** Call `list_council_options` and present the three with their
40
+ seat counts. A council is 7 to 44 seats and that range is the user's time and money.
41
+ **Skip the question when the request already answered it** — a named roster, "everything",
42
+ "be quick". A confirmation nobody needed is an interruption.
43
+ 2. Follow `skills/alphacouncil-agent/SKILL.md` from Stage 0. Do not improvise a shorter
44
+ workflow: a report that looks finished and skipped the bench is worse than an obviously
45
+ partial one.
46
+ 3. **Call the tools rather than searching for numbers they can supply.** `screen_ticker`,
47
+ `get_quote`, `get_options_chain`, `get_macro_snapshot`, `get_news`. Search is for
48
+ explanation and for what is not yet filed.
49
+ 4. Every selected master must report before the run is complete. `out_of_scope` is a
50
+ conclusion, not an abstention.
51
+ 5. Give price bands with the condition attached to each. "The cycle position is undetermined"
52
+ changes what the bands depend on; it does not excuse leaving them out.
53
+
54
+ ## Quick mode owes the reader one sentence
55
+
56
+ State in the output that no master lenses ran and nothing was cross-verified, so the
57
+ confidence is lower than a standard run. Without it, a four-seat run reads like a full
58
+ council result.
59
+
60
+ ## Screen mode
61
+
62
+ A rule whose inputs are missing is `skipped`, **never a pass** — reporting "6/7 passed"
63
+ without naming the seventh misrepresents the screen. Surviving is not a recommendation; it
64
+ means the name is worth research time. For a non-US ticker, say which market it is and use
65
+ `market_coverage` rather than returning nothing.
66
+
67
+ ## Market mode
68
+
69
+ Read it by the gap, not the ranking. A theme leading coverage while its series has not moved
70
+ means the story is ahead of the data, or the market stopped listening — say which. A series
71
+ that moved sharply while its theme is barely covered is the more valuable finding. Headline
72
+ counts measure attention, never truth.
73
+
74
+ ## Never
75
+
76
+ - Never fill a number from memory when a tool could have supplied it.
77
+ - Never let social or narrative evidence enter the conclusion on its own.
78
+
79
+ Educational software. The report says so.
package/AGENTS.md CHANGED
@@ -69,6 +69,32 @@ Verified against a real install (grok 0.2.101) rather than from documentation:
69
69
  Grok Build the workflow arrives through `AGENTS.md` plus the MCP tools, the same shape
70
70
  that works on OpenCode.
71
71
 
72
+ ## Slash commands
73
+
74
+ **One command, `/alpha`.** Modes are arguments, so there is one name to remember
75
+ rather than four in a menu of a hundred.
76
+
77
+ | Invocation | What runs | Model spend |
78
+ |---|---|---|
79
+ | `/alpha MU` | Full council — asks which preset first | one subagent per seat |
80
+ | `/alpha MU quick` | 4 analysts + debate, no bench, no verification | 7 seats |
81
+ | `/alpha MU screen` | Mechanical filings screen only | **none** |
82
+ | `/alpha MU options` | IV term structure, skew, positioning | **none** |
83
+ | `/alpha MU news` | Dated filings and headlines | **none** |
84
+ | `/alpha market AI` | What the market is talking about | **none** |
85
+ | `/alpha` | Lists the modes and stops | **none** |
86
+
87
+ The four marked **none** call keyless data tools and spawn no subagents, so they cost
88
+ nothing beyond the turn you type them in. The council modes spawn one subagent per seat, and
89
+ that is the entire cost of running this.
90
+
91
+ | Host | Where it reads them |
92
+ |---|---|
93
+ | Claude Code | `commands/` via `.claude-plugin/plugin.json`, plus `.claude/commands/` for a checkout |
94
+ | OpenCode | `.opencode/command/` |
95
+ | Grok Build | `.grok/commands/`, and `.claude/commands/` as a high-priority compatibility source |
96
+ | Codex | `~/.codex/prompts/` is **user-scoped**: `mkdir -p ~/.codex/prompts && cp commands/alpha.md ~/.codex/prompts/` |
97
+
72
98
  ## Market data coverage
73
99
 
74
100
  Structured financials come from each market's own regulator, and the pipeline degrades in
package/CHANGELOG.md CHANGED
@@ -2,6 +2,33 @@
2
2
 
3
3
  Notable changes per release. Dates are UTC.
4
4
 
5
+ ## [0.5.2] — 2026-07-26
6
+
7
+ ### Changed
8
+
9
+ - **Four slash commands collapse into one: `/alpha`**, with the modes as arguments.
10
+ `/alphacouncil-quick` and friends were long to type and, in a menu of a hundred commands,
11
+ four near-identical entries are harder to navigate than one. Modes: bare ticker for the
12
+ full council, plus `quick`, `screen`, `options`, `news`, and `market`. Invoked with no
13
+ arguments it lists the modes and stops rather than guessing.
14
+ - The modes that spawn no subagents are marked as such, because that is the actual choice a
15
+ user is making — `screen`, `options`, `news` and `market` call keyless data tools and cost
16
+ nothing beyond the turn.
17
+
18
+ ## [0.5.1] — 2026-07-26
19
+
20
+ ### Added
21
+
22
+ - **Slash commands**: `/alphacouncil`, `/alphacouncil-quick`, `/alphacouncil-screen` and
23
+ `/alphacouncil-market`, authored once and generated into every host's command directory.
24
+ 0.5.0 shipped with no command directory at all, so the plugin could only be reached by
25
+ @-mention or by describing it in prose.
26
+
27
+ ### Fixed
28
+
29
+ - The npm package did not carry the command directories, which meant an installed copy and
30
+ a checkout behaved differently.
31
+
5
32
  ## [0.5.0] — 2026-07-26
6
33
 
7
34
  Everything below is keyless: no API key, no account, no config file.
package/README.md CHANGED
@@ -123,6 +123,28 @@ The concise handoff is written to `~/.alphacouncil-agent/runs/<run_id>/user_resp
123
123
  The full report is written to `~/.alphacouncil-agent/runs/<run_id>/final_report.md`,
124
124
  with analyst Markdown files and `artifact_index.md` in the same run directory.
125
125
 
126
+ ### Slash commands
127
+
128
+ **One command, `/alpha`.** Modes are arguments, so there is one name to remember
129
+ rather than four in a menu of a hundred.
130
+
131
+ | Invocation | What runs | Model spend |
132
+ |---|---|---|
133
+ | `/alpha MU` | Full council — asks which preset first | one subagent per seat |
134
+ | `/alpha MU quick` | 4 analysts + debate, no bench, no verification | 7 seats |
135
+ | `/alpha MU screen` | Mechanical filings screen only | **none** |
136
+ | `/alpha MU options` | IV term structure, skew, positioning | **none** |
137
+ | `/alpha MU news` | Dated filings and headlines | **none** |
138
+ | `/alpha market AI` | What the market is talking about | **none** |
139
+ | `/alpha` | Lists the modes and stops | **none** |
140
+
141
+ The four marked **none** call keyless data tools and spawn no subagents, so they cost
142
+ nothing beyond the turn you type them in. The council modes spawn one subagent per seat, and
143
+ that is the entire cost of running this.
144
+
145
+ Available in Claude Code, OpenCode and Grok Build as soon as the plugin is installed. Codex keeps
146
+ its prompts user-scoped, so copy it once: `mkdir -p ~/.codex/prompts && cp commands/alpha.md ~/.codex/prompts/`
147
+
126
148
  ## What It Does
127
149
 
128
150
  Default stock-analysis runs are full runs, not lite summaries:
package/commands/alpha.md ADDED
@@ -0,0 +1,79 @@
1
+ ---
2
+ description: Equity research council — full run, quick read, mechanical screen, or market narrative
3
+ argument-hint: [ticker] [quick|screen|market|options|news] · or just a question
4
+ ---
5
+
6
+ # /alpha
7
+
8
+ Request: **$ARGUMENTS**
9
+
10
+ ## Route on the arguments — do not ask what they meant
11
+
12
+ | `$ARGUMENTS` looks like | Do this |
13
+ |---|---|
14
+ | *(empty)* | Print the mode table below and stop. Do not start a run. |
15
+ | a ticker alone, or a question | **Full council.** Go to "Full council" below. |
16
+ | ticker + `quick` | The `quick` preset: 4 analysts and the debate. No bench, no verification. |
17
+ | ticker + `screen` | `screen_ticker` only. No language-model judgment, no subagents. |
18
+ | ticker + `options` | `get_options_chain` only. |
19
+ | ticker + `news` | `get_news` on the symbol, and on its CIK when it is a US filer. |
20
+ | `market` (with or without a theme) | `get_market_narrative`. Add the theme as `extra_queries`. |
21
+
22
+ When `$ARGUMENTS` is empty, print exactly this and stop:
23
+
24
+ ```
25
+ /alpha MU full council — asks which preset first
26
+ /alpha MU quick 4 analysts + debate, no bench, no verification
27
+ /alpha MU screen mechanical filings screen only (no model spend)
28
+ /alpha MU options IV term structure, skew, positioning (no model spend)
29
+ /alpha MU news dated filings and headlines (no model spend)
30
+ /alpha market AI what the market is talking about (no model spend)
31
+ ```
32
+
33
+ Say plainly that the four marked modes call keyless data tools and spawn no subagents, so
34
+ they cost nothing beyond this turn. The council modes spawn one subagent per seat, and that
35
+ is where the spend is.
36
+
37
+ ## Full council
38
+
39
+ 1. **Ask which preset first.** Call `list_council_options` and present the three with their
40
+ seat counts. A council is 7 to 44 seats and that range is the user's time and money.
41
+ **Skip the question when the request already answered it** — a named roster, "everything",
42
+ "be quick". A confirmation nobody needed is an interruption.
43
+ 2. Follow `skills/alphacouncil-agent/SKILL.md` from Stage 0. Do not improvise a shorter
44
+ workflow: a report that looks finished and skipped the bench is worse than an obviously
45
+ partial one.
46
+ 3. **Call the tools rather than searching for numbers they can supply.** `screen_ticker`,
47
+ `get_quote`, `get_options_chain`, `get_macro_snapshot`, `get_news`. Search is for
48
+ explanation and for what is not yet filed.
49
+ 4. Every selected master must report before the run is complete. `out_of_scope` is a
50
+ conclusion, not an abstention.
51
+ 5. Give price bands with the condition attached to each. "The cycle position is undetermined"
52
+ changes what the bands depend on; it does not excuse leaving them out.
53
+
54
+ ## Quick mode owes the reader one sentence
55
+
56
+ State in the output that no master lenses ran and nothing was cross-verified, so the
57
+ confidence is lower than a standard run. Without it, a four-seat run reads like a full
58
+ council result.
59
+
60
+ ## Screen mode
61
+
62
+ A rule whose inputs are missing is `skipped`, **never a pass** — reporting "6/7 passed"
63
+ without naming the seventh misrepresents the screen. Surviving is not a recommendation; it
64
+ means the name is worth research time. For a non-US ticker, say which market it is and use
65
+ `market_coverage` rather than returning nothing.
66
+
67
+ ## Market mode
68
+
69
+ Read it by the gap, not the ranking. A theme leading coverage while its series has not moved
70
+ means the story is ahead of the data, or the market stopped listening — say which. A series
71
+ that moved sharply while its theme is barely covered is the more valuable finding. Headline
72
+ counts measure attention, never truth.
73
+
74
+ ## Never
75
+
76
+ - Never fill a number from memory when a tool could have supplied it.
77
+ - Never let social or narrative evidence enter the conclusion on its own.
78
+
79
+ Educational software. The report says so.
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "alphacouncil-agent",
3
- "version": "0.5.0",
3
+ "version": "0.5.2",
4
4
  "description": "Multi-agent public-equity research workflow plugin for Codex & Claude Code: sourced evidence packets, bull/bear debate, and a portfolio-manager Buy/Overweight/Hold/Underweight/Sell decision.",
5
5
  "type": "module",
6
6
  "license": "MIT",
@@ -56,6 +56,10 @@
56
56
  "LICENSE",
57
57
  "README.md",
58
58
  "README.zh-CN.md",
59
- "README.ja.md"
59
+ "README.ja.md",
60
+ "commands/",
61
+ ".claude/commands/",
62
+ ".opencode/command/",
63
+ ".grok/commands/"
60
64
  ]
61
65
  }